Skip to main content

Full text of "]In Search Of The Indo Europeans"

See other formats

In Search of the 

„ Indo 

In Search of the 

„ Indo- 


Language, Archaeology and Myth 

J.R Mallory 

With 175 illustrations 




The Discovery of the Indo-Europeans 

August Schleicher 14 The Indo-European hypothesis 22 

The Indo-Europeans in Asia 

The Anatolians 24 The Phrygians 30 
The Indo- Aryans 35 The Iranians 48 
Conclusions 63 

The Indo-Europeans in Europe 

The Greeks 66 The Thracians 72 The Illyrians 73 
The Slavs 76 The Baits 81 The Germans 84 Italy 87 
The Celts 95 Earlier configurations 107 

Proto-Indo-European Culture 

Environment 114 Economy 117 Settlement 120 
Technology 120 Social organization 122 Conclusion 126 

Indo-European Religion 

Dumezil and tripartition 130 Horse sacrifice 135 The cattle 
cycle 137 Human sacrifice and punishment 138 War of the 
Functions 139 Dualism and Indo-European ideology 140 
Mythology and reality 141 

The Indo-European Homeland Problem 

Defining the homeland 144 The neighbours of the Proto- 
Indo-Europeans 145 Internal linguistic evidence 151 
Interference and substrates 156 Linguistic palaeontology 158 
Archaeology 164 

The Armenians 33 
The Tocharians 56 

The Archaeology of the Proto-lndo-Europeans 186 

Dawn of the Proto-lndo-Europeans 187 Emergence of 
Proto-Indo-European society 188 The Eneolithic period of 
the western Pontic 195 The Early Eneolithic of the Pontic 
steppe and forest-steppe 197 Early Eneolithic in the 
East 206 The Yamnaya cultural-historical area 210 Proto- 
Indo-European culture 215 

Indo-European Expansions 222 

Expansion into Asia 223 Expansion into the Caucasus 231 
Southeastern Europe and western Anatolia 233 Central and 
Northern Europe 243 The process of expansion 257 
Recapitulation 262 


Epilogue 266 

The Aryan myth 266 The legacy 270 

Notes to the text 273 

Bibliography 278 

Sources of illustrations 284 

Acknowledgments 285 

Index 285 


By the first century ad historical records reveal peoples settled from the shores 
of the Atlantic to India all speaking languages closely related to one another. 
These are the Indo-European languages whose origins can be traced back to a 
common ancestor that was spoken in Eurasia some 6,000 years ago. We call the 
people who spoke this ancestral language the Indo-Europeans or Proto-Indo- 
Europeans. But although we can give them a name, they are unlike almost any 
other ancient people we are likely to encounter. As the linguistic ancestors of 
nearly half this planet's population they are one of the most important entities in 
the prehistoric record - and yet they are also one of the most elusive. No Proto- 
Indo-European text exists; their physical remains and material culture cannot 
be identified without extensive argument; and their geographical location has 
been the subject of a century and a half of intense yet inconclusive debate. 

To attempt to survey the origins of all the different Indo-Europeans and 
then track each of them to their original homeland and discuss their common 
culture is a task to daunt any single writer and certainly outrun the competence 
of any single scholar. Out in the academic world neither of these problems has 
ever been a serious deterrent, and in the past century there have been at least 
seventy volumes published as general surveys of the Indo-Europeans and their 
origins. Yet, other than sporadic attempts to resolve the problem of Indo- 
European origins with cursory reference to the different Indo-European 
peoples, there has not been a full general survey of the Indo-Europeans in 
English for at least a half century. This has encouraged me to produce this 
volume to fill the gap. 

During the course of writing this book, the authors of two recently published 
works on the Indo-Europeans were gracious enough to send me copies of their 
own books: Tomas Gamkrelidze and Vyachislav Ivanov's massive two-volume 
study Indo-European Language and Indo-Europeans (in Russian), and Colin 
Renfrew's more popular Archaeology and Language: The puzzle of Indo- 
European origins. I disagree rather fundamentally with both works which in 
some ways seem to have strayed light-years away from whatever consensus the 
general run of Indo-European studies has managed to achieve. Nevertheless, I 
have profited greatly from the vast accumulation of data offered by the two 
Soviet linguists and, while I am unconvinced of their solution to the problem of 
Indo-European origins, they will see in other discussions throughout this book 
the debt I owe their work. 

Colin Renfrew's book provided a stimulus of another kind. Although I had 
largely anticipated his conclusions on Indo-European origins in my original 


draft, my publishers encouraged me to make some additions to the basic text to 
take into account Professor Renfrew's most recent exposition of his theories. 
His latest work is very much a challenge to the conventional wisdom. While I 
regard my own work to be in the general mainstream of this 'conventional 
wisdom', I thought it out of place to reduce this book to an interminable 
counter-attack on a colleague's opinions. Consequently, I have examined 
Renfrew's main theories primarily in one part of my own text (Chapter Six), 
while reserving some more detailed argument for the notes. I should emphasize 
that I have not written this book simply to propose yet another solution to the 
Indo-European homeland problem; rather, I have also attempted to provide a 
general but, I hope, useful survey of the current state of our knowledge about 
the earliest Indo-Europeans. 

I believe that a discussion of the Indo-Europeans without the evidence of 
their languages would be like statistics without mathematics. For this reason, a 
number of linguistic 'figures' have been included in my belief that the general 
reader is far more interested in seeing what a line of Sanskrit or Gothic looked 
like than what pot the speakers of these languages may have cooked in or what 
device held their clothes together. Throughout this work I have always tried to 
keep in sight the fact that Indo-European is fundamentally a linguistic concept 
and that any cultural (pre)historian has certain obligations to the evidence of 
comparative linguistics. Nevertheless, I must plead guilty to both generaliza- 
tion and simplification. A certain graphic simplicity for linguistic forms is 
necessitated because too many diacritical marks, necessary though they may be 
for the proper articulation and analysis of the forms, have a way of terrifying a 
general reader. Those linguists who will immediately know what is missing 
from my forms will, I am sure, restore the vowel lengths, accents, and other 
necessary diacritics. I might add that any reference to Indo-Europeans in 
general, or to more specific Indo-European groups such as Greeks or Slavs, 
should be construed merely as short-hand for 'Indo-European-speaking' or 
'Greek-speaking' or 'Indo-European who occupied an area and later developed 
into a Greek-speaker', and no necessary reference to a specific physical type or 
material culture is intended. 

Although Indo-European is fundamentally a linguistic construct, I have 
written this book primarily from the perspective of an archaeologist who has 
been subjected to a certain number of the methods of the historical linguist. I 
have tried, as far as possible, to strike a balance between the evidence of the two 
disciplines, although I know only too well that the competing arguments for the 
'primacy' of archaeological or linguistic evidence will not satisfy everyone. 
Even with extended treatment, much of the archaeological discussion must, like 
the linguistic, be severely abbreviated to avoid losing both author and reader in 
incredible detail. As for the prehistoric (bc) dates cited in the text, these are all 
approximations based on the tree-ring-calibrated radiocarbon chronology, that 
is, bc = Cal. bc, or, for the reader unacquainted with these terms, bc dates are 
in ordinary calendar years. 


The Discovery of the Indo-Europeans 

My leisure hours, for some time past, have been 
employed in considering the striking affinity of 
the languages of Europe •; and finding, every day, 
new and most engaging entertainment in this 
pursuit, I was insensibly led on to attempt 
following them to their source. 

James Parsons, 1767 

That James Parsons approached his subject as a dilettante is obvious. Certainly 
his earlier studies on the human bladder, the structure of seeds, and 
hermaphroditism do not form the academic prelude that one might expect from 
someone seeking to trace the origins of the ancient peoples of Europe. But in 
fact, James Parsons, physician and fellow of both the Royal Society and the 
Society of Antiquaries, was probably no less well equipped to pursue such a 
study than any of his eighteenth-century contemporaries. The primary 
evidence for such an investigation was then limited to the more speculative 
efforts of ancient historians coupled with both pious and politically motivated 
fabrications of medieval monks, all of which was then constrained by a literal 
interpretation of the Book of Genesis. This confined all discussion to no earlier 
than 2350 b c (or about 1 ,656 years after the Creation) when the families of Noah 
and his sons disembarked from the Ark and set out to populate the world. The 
marriage of such diverse sources often required the eighteenth-century 
historian to find or forge correlations between the Bible and the Classical world 
resulting in such mammoth compendia as The Universal History from the 
earliest account of time to the present (1736-65). If Parsons had confined his 
investigations to these sources alone, his work could be justly dismissed as 
merely another academic curiosity presently disintegrating on a handful of 
library shelves. But Parsons recognized that there was a largely untapped 
source of evidence bearing on the most ancient peoples of Europe and Asia - a 
comparison of their different languages offered a guide to their relative affinity 
with one another and their distant origins. 

The close relationships between some European languages had already been 
clearly remarked upon by the beginning of the seventeenth century. Joseph 
Scaliger (1540- 1609), for example, attempted to divide the languages of Europe 
into four major groups, each labelled after their word for 4 god\ The transparent 
relationship of what we today call the Romance languages was recognized in the 
deus group (for example, Latin deus, Italian dio, Spanish dio, French dieu), and 
contrasted with the Germanic gott (English god, Dutch god> Swedish gud y and 


so on); Greek theos; and Slavic bog (such as Russian bog, Polish bog and Czech 
buh). Beyond this grouping Scaliger would not go, and he specifically denied 
any relationship between these different groups. However, during the course of 
the next century it became increasingly apparent to some that both the ancient 
languages and the peoples of Europe were more closely related than Scaliger 
had imagined. To those who preferred to take their historical evidence from the 
classical world, a wildly injudicious use of the term Scythian or Thracian came 
to be applied to most of those Europeans who had been situated north of the 
Greeks and Romans and who seemed to share some natural affinity. To those 
who preferred their history from the Bible, the label for these vaguely related 
Europeans was also easily obtained. Genesis had made it explicitly clear that the 
Semites (Jews, Arabs) and Hamites (Egyptians, Cushites) had derived from 
Shem and Ham respectively. It was then left to Noah's third son Japhet to 
father much of the remaining human race and hence it was not uncommon to 
lump the early peoples and languages of Europe under the name Japhetic. 

In 1767 Parsons published his study The Remains of Japhet, being historical 
enquiries into the affinity and origins of the European languages. Had this work 
been much shorter, its author might be better remembered. Unfortunately for 
Parsons, this rather tedious book ensured his obscurity and subsequent neglect 
in histories of Indo-European studies, a neglect not entirely deserved. 

Parsons began his linguistic survey by demonstrating the clear affinity 
between Irish and Welsh with an extensive (1,000 word) comparison of their 
vocabularies. This led him to the conclusion that Irish and Welsh 'were 
originally the same'. He then expanded his attention to the other languages of 
Eurasia by comparing their words for the basic numerals under the perfectly 
sound linguistic principle that 'numbers being convenient to every nation, their 
names were most likely to continue nearly the same, even though other pJrts of 
languages might be liable to change and alteration'. The comparisons were 
extensive and included Celtic (Irish, Welsh), Greek, Italic (Latin, Italian, 
Spanish, French), Germanic (German, Dutch, Swedish, Danish, Old English,' 
English), Slavic (Polish, Russian), Indie (Bengali) and Iranian (Persian). No 
one, no matter how untutored in the techniques of comparative philology, 
could fail to see similarities between the different languages in his list. In 
addition, in an exemplary instance of sound methodology, Parsons also listed 
the same numerals in Turkish, Hebrew, Malay and Chinese all of which failed 
to show any outstanding similarities either with the previous list of Eurasian 
languages or with one another. Parsons therefore concluded that the first group, 
the languages of Europe, Iran and India, were all derived from a common 
ancestor, the language of Japhet and his offspring, who had migrated out of 
Armenia, the final resting place of the Ark. 

In both proposing and demonstrating that the languages of Europe, Iran and 
India had all derived from a common ancestor, James Parsons could well be 
credited with having independently discovered what we now call the Indo- 
European language family. But Parsons shrouded his theory in a mass of 
biblical references, a gullible acceptance of the histories and chronicles of 



/ The outcome of Latin 
quattuor \four m various 
Romance languages shows how 
words for numerals tend to 
remain relatively stable 
a/though they experience 
phonetic change through time. 

2 Unlike numerals and other 
items of 'basic' vocabulary, 
most words are not so stable as 
can be seen in the various ways 
the Romance languages express 
the word for 'oak'. Some drew 
their word from Latin quercus 
'oak', specifically Quercus 
robur, others from the more 
general Latin robur 'oak, hard 
tree', while French retained an 
older Celtic form kassanos, 
parts of Iberia preserved a 
local word # kaxiku and 
Romanian adopted an old 
Balkan word gorun. 

medieval Irish monks, the mistaken inclusion of Hungarian among the related 
Japhetic languages as well as the assertion that North American Indian 
languages showed clear Japhetic characteristics. Finally, Parsons was guilty of 
the bizarre fallacy of Goropianism (after Goropius Becanus w ho had traced all 
languages back to Dutch) by assuming the pristine nature of Magogian (Irish) 
from whence all other Japhetic languages might be linguistically derived. 
Whether these mistakes coupled with the author's quite unrelated works on 
plant and human physiology sufficed to ensure his linguistic obscurity it is 
difficult to say, for the place of honour for the discovery of both the Indo- 
European family and comparative philology is traditionally assigned to Sir 
William Jones. 1 

In 1796 Jones, Chief Justice of India, founder of the Royal Asiatic Society, 
and, unlike Parsons, a scholar whose eminence in linguistic matters guaranteed 
the attention of the academic w orld, presented his famous discourse on Indian 
culture. During the course of the lecture, in what amounted to but little more 
than an aside, Jones made his famous pronouncement on the affinities of the 

1 1 
































Old English 































































































uZt e 2? k lfu nUmer f h ahrid S ed f'™3ames Parsons Its! and expanded to include 
Lithuanian, Albanian, Armenian and Tocharian. 

ancient language of India - Sanskrit - which I fear no historian of linguistics can 
resist quoting: 

The Sanskrit language, whatever may be its antiquity, is of wonderful structure; more perfect 
than the Greek, more copious than the Latin, and more exquisitely refined than either- yet 
bearing to both of them a stronger affinity, both in the roots of verbs and in the forms of grammar 
than could have been produced by accident; so strong that no phiiologer couid examine all the 
three without believing them to have sprung from some common source, which, perhaps, no 
longer exists. There is a similar reason, though not quite so forcible, for supposing that both the 
Gothic and Celtic, though blended with a different idiom, had the same origin with the Sanskrit; 
and the old Persian might be added to the same family. 

This model advanced by Jones suggesting a common and extinct ancestral 
language for the majority of the peoples of Europe, Iran and India has been seen 
by many as the first essentially modern exposition of the Indo-European 
theory. But perhaps this really places far too much credit on what Jones failed to 
convey to his audience in his brief lecture; we need only look to one of his later 
discourses to the same society to see how little Jones differed from Parsons 










































cii i 

























Tocharian A 
























































When engaging the problem of the 'common source' of these languages, Jones 
was content to follow the trail again back to the Ark whence issued the three 
great branches of humanity whose sons 'proceeded from Iran where they 
migrated at first in great colonies'. 

It is only in the first half of the nineteenth century that we see the actual 
development of a recognizable comparative philology and the growth of a 
concept of linguistic affinity unfathered by Noah. Rasmus Rask (1787-1832), 
for example, showed that it was not enough to allude to the intuitive linguistic 
similarity between various languages as was the practice of the earlier linguistic 
antiquarian; he argued that these similarities must be demonstrated systemati- 
cally. The affinity between the Greek word for 'oak', phegos, and English beech 
was founded on more than Japhetic intuition since it was predicated on a 
systematic correspondence of Greek ph = Germanic b; for example, Greek 
phero 'I carry' and English bear, or Greek phrater 'clan member' and English 
brother. Similarly, one could demonstrate the regular relationship between 
Greek g and Germanic k: Greek gyne, Old Norse kona 'woman'; Greek genos, 
Old Norse kyn 'family'; or Greek agros, Old Norse akr 'field'. 

In addition, it was not merely the similarities of sounds that were striking but 
the structure of the languages as well. The Sanskrit and Latin words for fire, 



Turkish Hebrew Malay 

1 bir e had 

2 iki 


satu yj 
s(a)nayim dua e r 
Uc salosa tiga 


4 dort arba'a 

5 bes hamissa 

6 alti 


empat si 



sissa enam |j u 

7 yedi sib'a 

8 sekiz 

9 dokuz tis'a 


tujoh qj 
semona (de) lapan ba 

sembilan jiu 

on asara su-puloh shi 

4 The basic numerals from Parsons four -non-Japhetic languages. 

agnis and ^respectively, are not only similar in sound but display similar 
changes in different grammatical cases: 

Sanskrit Latin 

Nominative Singular a gnis i gnis 

Accusative Singular a gnim ignem 

Dative/Ablative Plural agnibhyas jgnibus 

Such grammatical comparisons became the subject of major syntheses, the 
more famous of which were produced by Rask (1818) and Franz Bopp (,8i6 
1833). Rask continued the eighteenth-century tradition of ascribing to the 
ancestral speech an ethnic designation, in his case Thracian, but Bopp was 
content to leave the ancestral speech under the vague heading Stammsprache 
(original or source language), and the Book of Genesis began to evaporate 
nT m " St l7 uistic v disc ™.^ Indeed, as early as 1813 that remarkable 
polymath, Thomas Young, coined the term Indo-European in a review of 
Adelungs Mandates, a multi-volume attempt to discern the linguistic 
affinities of the world's languages by comparing translation texts of the Lord's 

August Schleicher 

By the mid-nineteenth century Indo-European studies were firmly established 
and major compendia of comparative philology were published. An excellent 
marker of the advances made by linguists of the time is the work of August 
Schleicher (i8 2I -r868) who provides a convenient point of departure for a 
number of topics. Schleicher was not only interested in systematizing the 
comparative evidence but also in elucidating the fundamental form of the Indo- 
European languages by working back through the linguistic history of each 
mdmdual language. In short, Schleicher set out to reconstruct the earliest 
Indo-European form of the words being compared. For example, before 







M y° enaean ^.Cyprian 

[.Arcadan , Atjic 





-1000 -500 

— Doric 






"Spoken Languages - 
(mid. Ind.) 

San p*r>c- 

- Phrygian 

L «an. 



B (Kuchean) 


S W.Pehlavi- Persian 
or Mid Pers 

1500 Modem Languages 

Modem Armenian 

Indo Aryan 

(Modem Indian Languages: 
Gujerati, Marathi. Panjabt. Hind. 
Hindustani. Bengali. Singhalese etc.) 




Modern Languages 






j— lllyriar . 


Dutch, Flemish. Fnsian 

5 The major Indo-European languages. 



Schleicher one might have only remarked the correspondence evident in the 
following list of words for 'field': 

Sanskrit ajras 

Greek agros 

Latin ager 

Gothic akrs 

But Schleicher could argue that in comparison with other Indo-European 
languages, it was clear that Gothic had regularly replaced a g with a k and that 
the vowel in the final syllable before s had been lost. The earlier Germanic form 
of Gothic akrs, therefore, must have been *agras. Similarly, in the mistaken 
belief that the Sanskrit language better preserved the Indo-European vowels, 
the underlying earlier form of Greek agros must have been *agras. Ultimately, a 
fundamental form *agras was postulated for the original Indo-European 
language and its outcome in each individual Indo-European language could be 
predicted according to each language's historical development. The recon- 
structed form of the word is marked with an asterisk to indicate that the word is 
not actually attested in any written source but is the product of linguistic 

The question as to what extent the reconstructions, or as some might prefer, 
linguistic triangulations, represent the 'original' language has always been a 
source of debate. There have been those who would argue that the 
reconstructed forms are founded on reasonably substantiated linguistic 
observations and that a linguist, projected back into the past, could make him or 
herself understood among the earlier speakers of a language. Others prefer to 
view the reconstructions as merely convenient formulas that express the 
linguistic histories of the various languages in the briefest possible manner. 
Their reality is not a subject of concern or interest. For those sceptical of claims 
to a practical reality for linguistic reconstructions, Schleicher's greatest folly 
was his attempt to write a folk tale in the reconstructed Indo-European 
language. He titled his tale 'Avis akvasas ka (The sheep and the horses). 

Schleicher's reconstructed forms were heavily influenced by a legacy that 
overvalued the utility of the early Sanskrit grammarians' description of their 
own language. By the twentieth century, comparative linguists had progres- 
sively altered the forms of their reconstructions and in the 1930s we find an 
updated rendition of the same tale, now Owis ek'woses-k">e by Herman Hirt. 
Today Hirt's reconstruction can be regarded as slightly archaic in light of more 
recent linguistic theory coupled with the evidence of the Indo-European 
languages of Anatolia, such as Hittite. As a gauge of the more recent changes 
some now prefer to reconstruct our unfortunate sheep as *H 3 owis or *Oewis, 
but as we are not concerned here with Indo-European phonology, the more 
traditional (and pronounceable) reconstructions of the early twentieth century 
will be employed throughout this work. 3 

Another of Schleicher's legacies was his model of the linguistic development 
of the Indo-European languages. Schleicher had always had a profound interest 



August Schleicher's version of 1868 

Avis akvasas ka 

Avis, jasmin varna na a ast, dadarka akvams, tarn, vagham garum 
vaghantam, tarn, bharam magham, tarn manum aku bharantam. Avis 
akvabhjams a vavakat: kard aghnutai mai vidanti manum akvams 

Akvasas a vavakant: krudhi avai, kard aghnutai vividvant- svas: 
manus patis varnam avisams karnauti svabhjam gharmam vastram 
avibhjams ka varna na asti. 

Tat kukruvants avis agram a bhugat. 

Herman Hirt's revised translation published in 1939 
Owis ekwoses-k w e 

Owis, jesmin w^lana ne est, dedork'e ekwons, torn, woghom 
g we rum weghontm, torn, bhorom megam, torn, gh'emonm ok'u 
bhertontm. Owis ekwomos ewe W ekw e t: kerd aghnutai moi widontei 
gh' e monm ek wons ag'ontm. 

Ek woses ewewekwont: k'ludhi, owei! k'erd aghnutai vidontmos: 
gh e mo, potis, weianam owjom kwrneuti sebhoi ghwermom westrom; 
owimos-k w e w e lsna ne esti. 

Tod k ek'ruwos owis ag'rom ebhuget. 

In 1979 Winfred Lehmann and Ladislav Zgusta published yet a third 
version with slight elaborations 

Owis ekwosk w e 

(G w arei) owis, kwesyo wlhna ne est, ekwons espeket, oinom ghe 
g*rum woghom weghontm, oinomkwe megam bhorom, oinomkwe 
ghmenm oku bherontm. 

Owis nu ekwobh(y)os ewewk^et: Ker aghnutoi moi ekwons agontm 
nerm widntei. 

Ekwos tu ewewk«ont: Kludhi, owei, ker aghnutoi nsmei 
widntbh(y)os: ner, potis, owiom r wlhnam sebhi gwhermom westrom 
k w rneuti. Neghi owiom wlhna esti. 

Tod kekluwos owis agrom ebhuget. 

A quite literal translation might run: 

[The] Sheep and [the] Horses 

[On a hill] [a] sheep, on which wool not was, saw horses, one, [a] 
wagon heavy pulling, [another] one, [a] load great, [another] one, [a] 
man swiftly carrying. [The] sheep to the horses said: heart pains me 
seeing [a] man horses driving. 

[The] horses to the sheep said: listen sheep, hearts pain us seeing: 
man, [the] master, wool of the sheep makes for himself [a] warm 
garment and to the sheep wool not is. 

That having heard, [the] sheep to the plain fled. 

A freer translation runs: 

The Sheep and the Horses 

[On a hill] asheepthat had no wool saw horses-one pulling a heavy 
wagon, another one a great load, and another swiftly carrying a man. 
The sheep said to the horses: it hurts me seeing a man driving horses. 

The horses said to the sheep: listen sheep! it hurts us seeing man, 
the master, making a warm garment for himself from the wool of a 
sheep when the sheep has no wool for itself. 

On hearing this the sheep fled into the plain. 

6 Three versions oj Schleicher's Proto-Indo-European tale. Each version employs a different 
Indo-European word j or the man mho is abusing sheep and horses alike. Schleicher reconstructed 
hisjrom the series: Sanskrit manus, Gothic manna, English man, Russian muz; Hirt drew his 
jrom the Proto-Indo-European *ghmon 'human, e.g., Latin homo, Gothic guma. Tocharian B 
saumo, Lithuanian zmuo; and the most recent version prefers *ner 'man, cf. Sanskrit nar- 
Avestan nar-, Greek ancr, Old Irish nert, etc. A fourth term, *wiros, means 'he-man 



in biology, and he employed the model of the genetic tree to describe the 
differentiation of the Indo-European languages. The Indo-European hypoth- 
esis presupposed that the great similarity between the various Indo-European 
languages could only be explained by assuming that they had all derived from a 
common language. This original language we would call Proto-Indo-European 
(or PIE) today. Over the area in which it was spoken, different regions began to 
diverge, branching off into various major language groups, or 'fundamentals', 
as Schleicher called them. Examples of the 'fundamentals' would be Celtic, 
Germanic or Slavic. By the same gradual process of divergence, these 
fundamentals split up into different languages, for example, Celtic (Irish, 
Welsh, Manx, Breton, and so on) or Slavic (such as Russian, Ukrainian and 
Polish), and these further branched into dialects or sub-dialects. In order to 
depict this historical process Schleicher resorted to the biological model of the 
family tree to describe the differentiation of the Indo-European languages. The 
length of the branches indicated the duration of time that various 
'fundamentals' had remained closely associated and the distances between 
branches indicated the degree of relationship. Thus, from Schleicher's diagram 
it was clear that the North European languages embraced by the Germanic, 
Baltic and Slavic languages were more closely related to one another than the 
South and West European languages. It could also be seen that the Germanic 
language had diverged from the northern branch long before Baltic and Slavic 
had separated. 

It is immediately apparent that the genetic model could not help but affect all 
discussion of the individual histories of the Indo-European languages and 
peoples. The model constrained one to frame questions such as when did the 
Germanic languages separate from the Balto-Slavic? And this was easily 
translated into the questions: When did the Germanic peoples move off from the 
Balto-Slavic peoples? When did the Celts separate from the Italic peoples? The 
Indians from the Iranians? And such historical questions demand a specific 
type of answer, not far removed from chasing after the offspring of Noah. 



7 August Schleicher s tree of 
the Indo-European languages. 



The overschematism of Schleicher's model of a family tree provoked a 
reaction. First, Schleicher's model had adopted the implicit premise that 
languages branch away from one another and remain isolated from contacts 
with other languages. But this w as surely contradicted by historical experience. 
Medieval French and Latin (Italic) has had an enormous impact on the English 
(Germanic) language. More importantly, Schleicher's model failed to explain 
many linguistic relations that were transparent even to a nineteenth-century 
linguist. For example, cutting across Schleicher's three main branches is the 
famous division between centum and satem languages. This refers to the Proto- 
Indo-European sound *k which has markedly different outcomes in different 
languages. The word for a hundred, *kmtom, yields the k sound in Latin 
centum, Old Irish cet, Greek hekaton, Gothic hund (from *kunt), but changes to a 
sibilant (s-sound) in Indie sata, Iranian satam, Lithuanian simtas, and so on. 
Schleicher's tree demanded an intensely close relationship between Germanic 
and Balto-Slavic, yet here Balto-Slavic were more closely allied with the Asian 
languages. For every readjustment of the branches, another linguistic isogloss 
(similarity) could be found to contradict it. 

Johannes Schmidt's (1843-1901) resolution of this problem was to go 
beyond a model of genetic development via ramification to one that imagined a 
broad band of Indo-European speakers in whose respective areas innovations 
developed and spread like waves to some, but seldom to all, other languages. 
The resulting image was therefore not a tree but a series of interlocked or 
encompassing circles that expressed specific similarities between one language 
and another. Schleicher's Balto-Slavic languages might share the same 
outcome of Proto-Indo-European but similarities with the Germanic 
languages were also depicted in the overlapping circle. This provided a 
somewhat more realistic presentation of the linguistic relationships between 
languages but it also rested on a major premise: the positionings of the 
languages in the diagram portrayed geographical realities without specifying 
the time at which anything occurred. The model, in short, was a synchronic 

8 Johannes Schmidt's wave 
mode! of the Indo-European 
languages. Those languages 
encompassed under I all share 
an e where Indo-Iranian has 
an a, e.g., Latin est but 
Sanskrit asti. Those in II 
change *k to an s-sound 
(centum versus satem), while 
those in group III form some 
case endings in m rather than 



description of the Indo-European languages but offered little or no historical 
perspective. The Celts were simply people who spoke a language that seemed to 
share certain similarities with the Italic languages. That some of these 
similarities may have been inherited from the original Proto-Indo-European at 
one period, and that others were only acquired perhaps thousands of years later, 
could not be discerned in Schmidt's diagram. To be fair to Schmidt, it was not 
his intention to replace Schleicher's model but rather to provide another 
(synchronic) dimension to the interrelationships of the Indo-European 
languages. A more subtle tool than Schleicher's, Schmidt's wave model still 
failed to convey the historical development of the languages. 

While we might wish to turn now to a universally accepted model of the 
development of the Indo-European languages, unfortunately, all subsequent 
attempts to depict the interrelationships of the Indo-European languages 
invariably collapse into the two dimensionality of Schleicher or Schmidt, albeit 
updated with far more information and vastly more complex articulations. The 
modern historical linguist will still find it hard to escape from the use of the 
'fundamentals' of Schleicher or separate out the complex stratigraphy of Indo- 
European isoglosses that make up a wave model. There is, to be sure, 
unequivocal agreement on some matters, for example, that the Indians and 
Iranians were extremely closely related (linguistically) before their emergence 
into the historical record; yet Indo-European dialectology still remains a 
fruitful and frightful area of research and debate. 

Phase l 

Northern Group 

Phase II \ West IE -Slavic -Baltic 


\Greek — Thracian — Armenian — Aryan 
Northern Group 

Southern Group 

pnase |(l s West jE - Slavic - Baltic 

Celtic i italic 

Southern Group 

g The development of the Indo-European languages according to Francesco Adrados (1982). 



10 The development of the Indo-European languages according to Tomas Gamkrelidze and 
Vyacheslav Ivanov (1985). 

18 N 

// A modern 'wave model' 
of the Indo-European 
languages according to Raima 
Anttila (1972). The numbers 
indicate 24 isoglosses 
(similarities) shared among 
different Indo-European 
languages. Isogloss 1 indicates 
the centum.satem split 
(Schmidt's wave II). A 
number of isoglosses, generally 
interpreted as innovations, 
appear to link Greek-Iranian- 
Indic-A rmenian; simila rly, 
Germanu-Baltic-Sla vie sha re 
a number of similarities, while 
a large isogloss 'bundle' divides 
Italic from Greek. I I it t it e and 
Tocharian tend to be on the 
conservative side of most 
dialect developments in 
accordance with their peripheral 
positions with respect to the 
other Indo-European 

12 16 



The Indo-European hypothesis 

If details are still a matter of controversy, the Indo-European hypothesis is 
most certainly not. It is the only explanation that can convincingly account for 
why approximately half the earth's population speaks in languages clearly 
related to one another. This requires the assumption that at some time and some 
place in Eurasia there existed a population which spoke a language directly 
ancestral to all of those we now recognize as Indo-European. The usual label for 
such a language, Proto-Indo-European, is also applied to speakers of this 
language who are the subject of this book. Furthermore, it is assumed that they 
once occupied an area considerably more circumscribed than that where we 
find them when they first enter the historical record. Unfortunately, this 
conclusion has all too often rested on a form of latent Japhetism - a tendency of 
nineteenth- and even twentieth-century writers to portray the Proto-Indo- 
Europeans as a single people constrained within their homeland, perfecting 
their language and then bursting out all over the earth waving swords and 
spreading paradigms. This metaphor haunts much of what has been written 
about the Indo-Europeans as a people. But its failure wholly to convince, or 
worse, its ability to attract derision rather than critical appraisal, has little to do 
with the validity of our concept of a confined homeland and people. Since this 
element of our perception of the Indo-Europeans is fundamental to much of 
our investigations here, we need to consider this aspect more closely. 

The one constancy of language is that it is always changing; it is stasis that is 
unnatural since this would require numerous individuals over a period of 
generations to reproduce precisely the same sounds and replicate the same 
idiom, a task which is contrary both to the nature of human behaviour and to the 
necessity for language to accommodate continual change in culture. The 
complex reasons for language change - for example, social class, phonetic drift, 
analogy, contact with other languages - cannot concern us here. We merely 

Modern English 

Yahweh is my shepherd, I lack nothing. 
In meadows of green grass he lets me lie. 
To the waters of repose he leads me. 

Early Modern English 

The Lord is my shepherd, I shall not want. 
He maketh me to lie down in green pastures. 
He leadeth me beside the still waters. 

Middle English 

Our Lord gouerneth me, and nothyng shal defailen to me. 
In the sted of pasture he sett me ther. 
He norissed me upon water of fyllyng. 

Old English 

Drihten me raet, ne byth me nanes godes wan. 

And he me geset on swythe good feohland. ' 2 Translations of the 2 3 rd 

And fedde me be waetera stathum. Psalm illustrate how the 

English language has changed 
1 over the past /,ooo years. 



The Early Distribution of the 

Indo-European Languages 




ij The early historic al distribution of the major Indo-European linguistic groups. 

need contemplate that the English language spoken by Chaucer would be on the 
outer reaches of intelligibility to a speaker of twentieth-century English. And 
when confronted with the Old English language, the modern English speaker 
swiftly bids goodbye to comprehension in the same way that a Frenchman, 
Spaniard or Romanian would despair of carrying on a conversation with an 
ancient Roman. If continuity of language can be ascribed it is not because of 
stasis but rather because its speakers, over time and in a specific region, have 
maintained a course of relatively similar linguistic change. 

It should be clear then that several factors severely affect linguistic 
continuity, or parallel language change. We may expect that the amount of 
change will be partly dependent on the extent of time that has elapsed in the 
linguistic continuum. Also, it will be affected by the size and nature of the 
geographical area occupied. In the absence of mass media or a written standard, 
people speaking originally the same language but separated by large distances 
are unlikely to maintain parallel changes. Where we find great similarity of 
speech over a large area we can normally assume a recent expansion since the 
factors of time and distance will normally reduce a single language into a 
continuum of mutually related but increasingly different languages. This being 
so, the similarity of the Indo-European languages when we first encounter 
them historically, stretching from the Atlantic to India, all speak for their 
relatively recent spread from a more confined area. To ascribe to such a 
dispersion a very great antiquity would be to attribute to the Proto-Indo- 
European language properties w holly contrary not only to all the evidence from 
the world's other languages but also to human behaviour itself. In short, the 
Indo-European hypothesis presupposes a Proto-Indo-European language 
spoken by a population in some area of Eurasia severely more confined than 
their earliest historical distribution. How confined, when and where will be 
problems for later chapters. 



The Indo-Europeans in Asia 

Asia is called the cradle of mankind. The times 
change. People change. But the belief in the 
Asiatic cradle has not changed. 

Hans von Wolzogen, 1875 

The Indo-Europeans did not burst into history; they straggled in over a period 
°f 3>5<>o years, announcing their arrival in the historical record in as varied 
media as clay tablets in Anatolia and Greece, inscriptions carved on the face of 
an Iranian cliff, a dedicatory inscription on a German helmet or a Lutheran 
catechism for pagan Lithuanians. But no matter when or how we first 
encounter the language of Indo-European speakers, they all have one thing in 
common: they invariably speak an already differentiated Indo-European 
language, never Proto-Indo-European. As this work seeks to trace the origins of 
the earliest Indo-Europeans, we must confine our attention to that brief 
moment when each Indo-European group first emerges in the historical record, 
and then seek its more immediate linguistic and archaeological origins. In so 
doing, we hope to draw nearer to the ancestral Indo-European community 
whence they all originated. 

We cannot, however, pursue this course without making reference to some 
fundamental archaeological terms. These will include 'cultures', which are 
traditionally defined as the recurrence of similar ceramics, tools, architecture 
and burial rites over a limited area, for example, the Gandhara Grave culture of 
the Swat Valley in Pakistan. It should be emphasized that such cultures are the 
constructs of modern archaeologists and their correspondence with actual 
prehistoric social groups is a constantly debated topic. The evidence for these 
cultures is to be found in archaeological sites which may contain various levels 
dating to different periods, each designated with a numeral, such as Troy V (the 
fifth major phase of the site of Troy). Events such as the destructions of sites, or 
the appearance of a particular item of culture, may be found across various 
cultures at apparently the same time and constitute an 'horizon'. 

Since the historical record itself begins in Asia, this is where we will first take 
up their trail. 

The Anatolians 

The earliest Indo-European-speaking peoples to enter the historical record 
were the Anatolians who are first attested by about the nineteenth century bc. 
By this time Assyrian merchants had penetrated into south central Anatolia and 


established their karum or 'trading office' at Kanes, the modern Kiiltepe. 
Excavations at this site, and at several other Assyrian trading posts, have 
uncovered clay tablets in Assyrian cuneiform that record the daily business of 
the Assyrian tradesmen. In addition, they also mention personal names and 
places which are recognizably Indo-European. 4 By about the mid-seventeenth 
century bc the Indo-European speakers are declaring themselves in several 
different Anatolian languages. By far the best attested of these is Hittite. With 
their capital at Hattusa (modern Bogazkoy), the Hittites have left us over 25,000 
clay tablets spanning the period from about 1650 to 1200 bc. In addition, their 
archives contain tablets in two other Indo-European languages, Luwian and 
Palaic. They present us with a picture of Anatolia where the Hittites are masters 
of the central region, the Palaic speakers subservient to their north, and the 
Luwians occupying the role of traditional rival in much of western and southern 
Anatolia. After the collapse of the Hittites about 1200 bc, Luwian seems to have 
prevailed widely over southern Anatolia and Luwian-related languages such as 
Lycian continued down into the last centuries bc, only to be finally engulfed by 
the expansion of Greek colonists. 

It is most important for our purposes to inquire how autochthonous were the 
Anatolian languages in their respective regions. The general opinion of both 
linguists and archaeologists would almost universally deny them a role as 
natives to Anatolia but cast them rather in the part of Bronze Age intruders who 
assimilated the indigenous non-Indo-European populations. The Assyrian 
merchants of the nineteenth century bc not only record the names of Indo- 
European peoples in their texts but make it quite clear that there was also a great 
body of non-Indo-European-speaking peoples in the region. The existence of 

14 The curliest Indo-European languages of Anatolia and their non-Indo-European neighbours 
(in upper ease). 



these non-Indo-European peoples is undoubted since the Hittite archives 
themselves contain texts, translations of texts, and frequent borrowings from a 
language called Hattic. These Hatti are regarded as the predominant 
substratum, the aborigines if you will, of central Anatolia over whom the 
Hittites and Palaic speakers superimposed themselves. From the Hatti the 
Hittites borrowed not only many words, 5 but also much of their culture, 
certainly much of their religion, and even the name Hittite derives from Hatti 
(the Hittites called themselves ties and their language nesili). Linguistically, 
Hattic is a non-Indo-European language with no certain close relationships, 
although there are some grounds (absence of grammatical gender, use of 
prefixes) to link it with the northwest Caucasian group of languages (Abkhaz) or 
perhaps Kartvelian, the major south Caucasian linguistic group. 

Further to the east, on the fringes of Anatolia and north Syria, lay another 
major non-Indo-European people, the Hurrians. Hurrian texts maintained in 
the Hittite archives, coupled with Hurrian loan words in Luwian and the 
Hurrians' own inscriptions and texts in north Mesopotamia which date as early 
as the twenty-third century bc, all speak for an additional non-Indo-European 
presence on the eastern borders of the Indo-Europeans of Anatolia. To their 
south were the lands of the Semites and (formerly) Sumerians, again non-Indo- 
European speakers. The natural conclusion to be drawn from all of this is that 
the Indo-European-speaking Anatolians were intrusive into central Anatolia 
and were unlikely to have emigrated from directly east or southeast of this 
region where major non-Indo-European populations are historically attested. 
It is also clear from the abundance of mixed texts, foreign loanwords in Hittite 
and Luwian, and the entire cultural picture that emerges from the content of the 
texts, that the Indo-European Anatolians had already undergone considerable 
assimilation to the culture of the non-Indo-European Anatolians before they 
appear in history. Now what do linguistics and archaeology tell us about their 

For the linguist, the existence of three Indo-European languages in Anatolia 
by the seventeenth century bc generates two issues of considerable historical 
importance. The first is their relationship with the other Indo-European 
languages. Here there is fairly universal agreement among historical linguists 
that the Anatolian branch offers us some of the most extreme examples of 
archaism among all the Indo-European languages. By this is meant that they 



n = asta assu 

malhip = hu 

Then goodness 

! anda tarneskiddu 


should he let in. 

j idalu = ma = kan 

asah = pi 

But evil 

anda le tarnai 


should he not let in. 



(The god) Sulinkatte, 



the king. 

anda eszi. 

a-ta-niua a s 

sits within. 

IS A bilingual religious text in Hittite and Hattic. 
The Hittite text contains Indo-European words, e.g., 
assu 'good' is related to Sanskrit su, Greek eu 'good'; 
Hittite anda 'within is cognate with Latin endo; and 
eszi 'sits* is to be compared with Sanskrit aste and 
Greek asti. Sultnkattis, however, is a Hattic name 
which the Hattic text reveals to be made from 
Suhn + katti 'king. LVGAL indicates the Sumenan 
sign Jo r ' king ' . Stru ctura //) ' Hittite a nd Ha i t ic a re 
very different languages, the Hittites borrowed much 
of their religion from the indigenous Hatti. 



1 6 Some basic numerals and 
kinship terms indicate the 
fundamental differences 
between Hurrian and Indo- 

nun loi i 

Prntn-lnrlrvFi irnnpan 
nuiu ii iuu lui i 

















*bh rater 




retain grammatical forms and constructions that disappeared very early on in 
the other languages. 6 Some would go further and argue that the Anatolian 
branch appears to lack some grammatical forms that developed in all the other 
Indo-European languages. This, they maintain, indicates that the Anatolian 
branch diverged from the rest of the Indo-European continuum before it had 
even evolved into the form of Proto-Indo-European that gave rise to all the 
other Indo-European languages. This view is not universally accepted, 
especially since most linguists admit that the Anatolian languages had already 
undergone vast changes under the influence of non-Indo-European native 
populations before they emerged into history. Although there is much finely 
argued controversy about the details of these linguistic issues, there would be 
few to argue against the conclusion that the Anatolian languages represent a 
very early separation or divergence from the common Proto-Indo-European 
continuum of dialects. 

The second major issue is the internal relationship among the Anatolian 
languages. With our evidence for both Luwian and Palaic so meagre compared 
with Hittite, it is difficult to ascertain fully how divergent the three languages 
were. That differences did exist can easily be seen in comparing some of their 



















But despite these differences, and some are more thoroughgoing than these, the 
three languages are vastly more similar to one another than they are to any of the 
other Indo-European languages, even those other languages which are also 
attested as early as the Bronze Age. They give all the appearance of being the 
result of linguistic differentiation across a broad band of common Anatolian 
dialects. Their divergence from one another must obviously have occurred 
before their earliest historical attestation, but not too long before or we would 
expect yet greater differences. Linguists normally provide a broad estimate that 
the ancestors of the different Anatolian languages penetrated into their 
respective territories some time during the third millennium BC, or possibly as 



early as the later fourth millennium. Where does all this leave the archaeologist? 

First, the Indo-European-speaking Anatolians are difficult to distinguish 
from their non-Indo-European neighbours or predecessors. They appear to 
have embraced thoroughly the local Anatolian Bronze Age cultures and they 
display no obvious cultural traits that mark them off as distinctly Indo- 
European. This is hardly surprising, as the basic social picture of Bronze Age 
Anatolia is of a series of city-states comprised of linguistically diverse 
populations sharing the same material culture. It has even been suggested that 
Hittite itself was not the language of the dominant group but rather a lingua 
franca, developed out of the close association of the earlier Hittites of Kanes 
with the Assyrian merchants, who were the first literate population in Anatolia 
and who used Kanes as a trading base. 7 

We must also remember that our knowledge of Anatolian archaeology is still 
quite inferior to many other areas of Eurasia and so any arguments for ethnic 
intrusions are generally built on admittedly meagre evidence. This is more than 
compounded by the length of rope with which the linguist has provided the 
optimistic archaeologist, because with a i ,500-year time span to seek intrusions, 
few archaeologists who believe that such phenomena are traceable in the 
archaeological record can resist discovering several possible invaders - from 
both the west and the east. 

Probably the most widely accepted case for intrusion falls at the end of Early 
Bronze Age II, about 2700-2600 BC, when the evidence for population 
movement is coupled with destruction and abandonment. Beginning in western 
Anatolia we see destruction phases on every major site and the abandonment of 
smaller sites. The Konya Plain is offered as the most convincing example since 
field surveys here have indicated a collapse from 100 Early Bronze Age II sites 
to a mere four in the following period. Some suggest that an infiltration by 
nomads who profoundly altered the sedentary economy of the region may be 
credited with this change. In addition, new ceramic elements which take their 
origin from northwest Anatolia (Troy V) spread rapidly eastward as also does 

// Principal sites associated with theories of Anatolian invasions. 



the classic form of status or ritual architecture - the megaron which was 
common at Troy and Beycesultan - which now begins to appear in central 
Anatolia at such sites as Kanes-Kultepe. 

The arguments for a west to east movement of intruders in the mid-third 
millennium BC accords well with some linguistic theories concerning the 
dispersion of the Anatolian languages. Essentially, the new horizon embraces 
the subsequent historical lands of the Luwians who maintained a west to east 
pressure throughout their existence. This crisis at the end of Early Bronze Age 
II may have been either the manifestation of the earliest Luwians or even the 
earliest Anatolian speakers, including the ancestors of the Hittites, who 
underwent subsequent linguistic differentiation. The abandonment and 
destructions may nevertheless have simply been the result of climatic or 
internal calamities while the spread of yet another ceramic style or architectural 
form may not have required a new people with a new language. 

The original nucleus of these proposed expansions is northwest Anatolia, 
which naturally includes Troy itself. Links between this region and Southeast 
Europe, especially in ceramics - including figurines - and architecture, have 
long been known, and until the past few decades generally attributed to an 
expansion of Near Eastern high culture to European barbarians. More recently 
there has been a recognition by some archaeologists that the direction of 
influence may require reversing, at least during the transition from the 
Chalcolithic to the Early Bronze Age. This can be seen, for example, in the 
ceramics, metallurgy and architecture exhibited at Bulgarian sites such as Ezero 
which only appear later at Troy. These similarities are seen by some to be little 
more than a general cultural horizon embracing both sides of the Sea of 
Marmara while others argue for actual folk movements, possibly refugees, who 
abandoned the Balkans for northwest Anatolia about 3500—3000 BC under 
either the pressure or leadership of the Indo-Europeans. The remains of horse - 
whether wild or domestic is not certain - at Anatolian sites such as Demirci 
Huyiik is also cited as evidence for intrusions from Southeast Europe where the 
domestic horse antedates the Anatolian evidence and is a known possession of 
the earliest Indo-Europeans. We are not yet prepared to follow such a trail so 
early in our enquiry since this concerns too closely the problem of the Indo- 
European homeland itself. Rather, we must briefly turn our attention to those 
who prefer to seek the origins of the Indo-European-speaking Anatolians to the 

Most arguments for an Indo-European invasion from the northeast concern 
the appearance of a new burial rite at the end of the fourth and through the third 
millennium bc. At this time, both north of the Black Sea and the Caucasus, 
burials on the Russian-Ukrainian steppe were typically placed in an under- 
ground shaft and covered with a mound (kurgan in Russian). Before 3000 BC 
there begin to appear in the territory of the indigenous Transcaucasian (Kuro- 
Araxes) culture somewhat similar burials such as the royal tomb of Uch-Tepe 
on the Milska steppe. As tumulus burials are previously unknown in this 
region, some would explain their appearance by an intrusion of steppe 



pastoralists who migrated through the Caucasus and subjugated the local Early 
Bronze Age culture. More importantly, a status burial inserted into a mound at 
the site of Korucu Tepe in eastern Anatolia has been compared with somewhat 
similar burials both in the Caucasus and the Russian steppe. The discovery of 
horse bones on several sites of east Anatolia such as Norsun Tepe and Tepecik 
are seen to confirm a steppe intrusion since, as mentioned earlier, the horse, 
long known in the Ukraine and south Russia, is not attested in Anatolia prior to 
the Bronze Age. 8 Continuing contacts or migrations are employed to explain 
subsequent similarities between the royal tombs of north-central Anatolia, such 
as the thirteen graves of Alaca Hiiyiik, with tombs formally similar or 
possessing related grave goods known north of the Caucasus. 

At present, a northeastern intrusion does not make quite so good a linguistic 
'fit' as does the northwestern hypothesis. The evidence for intrusion is either 
confined to eastern Anatolia - lands historically attributed to Hurrian or 
Caucasian languages - or north-central Anatolia where we might expect Hattic 
or Kaskian, another apparently non-Indo-European linguistic group. The 
evidence of kurgan-related burials is generally absent from those territories 
where we find the major Indo-European peoples of our search, especially the 
Luwians of southern and western Anatolia. And even if our kurgan-entombed 
overlords are proximate to the Hittite's traditional territory, linguists ada- 
mantly oppose an eastern entry for the Hittites and a separate western entry for 
the Luwians. The languages seem too closely related, too similar to have 
experienced the degree of separation implied by each having taken an opposing 
course around the Black Sea. Furthermore, what similarity exists between royal 
burials both north and south of the Caucasus may have far more to do with the 
need to develop more impressive forms of entombment for the hierarchies that 
developed in both regions during the Early Bronze Age, and which participated 
in mutual exchange networks of prestigious goods. At present, the scales are 
tipped in favour of a western entry. 

The Phrygians 

With the collapse of the Hittite empire about 1200 bc, the configuration of 
Indo-European languages in Anatolia changes. Luwian speakers continued to 
dominate southern Anatolia and parts of north Syria and were so successful at 
adopting the mantle of the Hittites that they retained this ethnic designation, at 
least in the Bible. These Luwians employed a hieroglyphic script which due to 
its discovery in the Hittite capital formerly bore the mistaken name of 
'Hieroglyphic Hittite'; now it is more commonly called Hieroglyphic Luwian. 
As we have seen, in western Anatolia other late Anatolian languages such as the 
Luwian-derived Lycian continued until its speakers were assimilated by Greek 
colonists. But in central Anatolia itself, in the heartland of the Hittite empire, 
there appeared a new Indo-European language, Phrygian, which stands outside 
the Anatolian group and cannot be derived from any of the former Indo- 
European languages of Anatolia. 



18 Anatolia after the collapse of the Hittites. Indo-European languages are indicated in 
italics; non-Indo-European languages are in upper case. 

From 1200 bc to 800 bc the Phrygians are commonly associated with other 
intruders who swept across central Anatolia from the west, participating in the 
destruction of the major Hittite cities and plunging Anatolia into its 'Dark Age'. 
The archaeological evidence for these invaders is not yet secure, especially since 
there is an embarrassing gap between their occupation of Troy Vllb in the 
twelfth century and the earliest emergence of distinctive Phrygian pottery in 
central Anatolia during the eighth century bc. Their handmade black ware is 
known from numerous sites, including the ancient Hittite capital at Hattusa 
and the Phrygian's own capital of Gordion. As a power they dominated central 
Anatolia for several centuries and produced a number of figures of both Greek 
legend and Anatolian history, who- all bore the name of Midas. 

The historic King Midas ruled from the eighth century B c, and in his contest 
with the Assyrians and his attempt to expand the frontiers of Phrygian power, 
he brought the state to its apex, only to see it collapse under the pressure of 
Kimmerian invaders from the north (who drove Midas into committing suicide 
by drinking ox-l?lood). More famous, naturally, is the legendary Midas whose 
story is variously set in both Macedonia and Phrygia, a confusion which we will 
soon see dovetails neatly with the whole problem of Phrygian origins. Having 
captured a silenus - an older, wiser but even more besotted version of the 
classical satyr - by spiking a spring with w ine, Midas returned him to Dionysus 
and was rewarded with his fondest wish - the golden touch. When the starving 
Midas discovered that even his food and drink turned to gold on touch, he 
repented his folly to the god and was instructed to wash himself in the river 
Pactolus. This river instantly became a major source of gold and provided the 
economic foundation for that other great symbol of wealth, Croesus. The 
wealth of the Phrygians was not purely legend since one of the relicts attributed 
to them are their large and rich tumulus burials. 

Linguistically, we are severely limited by the sparse remains of the Phrygian 
language. The earlier inscriptions number only about twenty-five and date to 
the period 800-600 BC, while about 100 much later inscriptions, generally 



Old Phrygian 

Ates arkia evais akenan o-lavos Midai lavaltaei vanaktei edaes. 

Ates? dedicated and carved this stone for Midas, the protector of 
the people, the king. 

Late Phrygian 

ios ni semoun knoumanei kakoun addaket, 

gegreimenan egedou tios outan 

akke oi bekos akkalos tidregroun eitou 

Who does evil to this grave 
he bears the inescapable curse of god 
and to him shall bread and water be unpalatable' 

/o Two Phrygian inscriptions. The later inscription contains the word kakoun 'evil which 
may be compared with Greek kakos l bad\ lent into English in words like cacophony. The word 
would appear to derive from a Proto-Indo-European child's word *kakka which should not 
require translation, cf. Armenian kakor *excrement\ Greek kakkao, Latin caco, Middle Irish 
caccaim, etc. At the other end of the spectrum is tios which some have attempted to relate to 
Sanskrit dyaus, Greek Zeus and similar words for the Indo-European sky-god. Note also that 
the later inscription employs bekos which we know from Herodotus to mean * bread'. 

unsuccessful interdictions against tomb plundering, are recorded from the first 
centuries ad. In addition, we have the customary evidence of early place names 
and personal names and a few glosses (marginal comments in manuscripts 
providing definitions for otherwise obscure words). One of these comes to us in 
an entertaining story by Herodotus, who relates how the Egyptian king 
Psammetichus (663-609 bc) once attempted to discover the oldest language in 
the world. This was accomplished by ensuring that two infants heard no human 
speech until they had uttered their first words, under the assumption that these 
would be from the primeval language. The children's first word was bekos 
which Herodotus informs us was the Phrygian word for bread, and hence 
Phrygian was reputed to be the oldest language in the world. This fantastic 
story may do little to encourage us to accept Herodotus as a thoroughly reliable 
source when he also tells us how the Phrygians originally lived in Macedonia 
and migrated into northwest Anatolia, perhaps about the time of the Trojan 
War, where they changed their name from Bryges to Phrygians. Nevertheless, 
consensus does derive the Phrygians from Southeast Europe. 

Although the sparseness of the linguistic evidence makes it difficult to assess 
the dialectal position of Phrygian within the Indo-European languages, there is 
certainly no convincing evidence that it was specifically an Anatolian language 
closely related to Hittite, Luwian or Palaic, and hence there are sound linguistic 
grounds to regard it as a later intruder. There have been a number of attempts to 
link Phrygian closely with Thracian and Illyrian, two major Indo-European 
languages of the Balkans which offer even more appallingly sparse linguistic 
remains, and many handbooks speak of the Thraco-Phrygian group of 
languages. Today such a close relationship is still unproven, although Phrygian 
has more affinities with the Balkan languages than with any others. 



Consequently, it is from that direction that the linguist seeks the original 
Phrygians and defers to Herodotus's account as at least more plausible than any 
alternative hypothesis. 9 

The historical and archaeological evidence is much too meagre to erect a 
convincing case for Phrygian origins. It is perhaps ironic that the best 
archaeological evidence for an intrusion, the abrupt break between Troy Vila 
and Vllb with its introduction of Southeast European Knobbed Ware about 
1200 bc, although fitting remarkably well with Herodotus's account, gets us no 
further than northwest Anatolia. Consequently, it is difficult to employ this 
evidence to support the late Hittite records of a King Midash (an early Phrygian 
Midas?) on its northern frontier prior to the destruction of Hattusa. It is only in 
the eighth century bc, when the inscriptional evidence and a distinctive pottery 
style cojoin to provide us with an unambiguous Phrygian presence, that we can 
safely mark their existence in Anatolia. The one clearly intrusive item of their 
culture is the tumulus burial, especially prominent at Gordion. Large tumuli 
are well known in the Balkans and have a much longer ancestry there, 
consequently, a link between the two regions on these grounds has long been 
suggested. But since such tumuli only appear in Phrygian territory from the 
eighth century onwards, the link between a Balkan homeland and Phrygia can 
only be entertained if we detach the Phrygians from the destruction horizon 
associated with Troy VII and Hattusa. Before attempting any conclusions here, 
we should first look at one more group of intrusive Indo-Europeans, the 

The Armenians 

In his enumeration of the great army of the Persian king Xerxes (519-465 bc), 
Herodotus includes the Armenians, whom he informs us were 'Phrygian 
colonists'. On the basis of this statement, coupled with the linguistic position of 
Armenian within the Indo-European family, it is generally accepted that the 
Armenians, like the Phrygians, emigrated from the Balkans into Asia Minor. 10 
However, unlike the Phrygians whose language has long been extinct, 
Armenian is still spoken today by approximately five million people. Its earliest 
texts, largely religious, are traditionally ascribed to the fifth century ad but are 
probably more recent. Nevertheless, there are no grounds whatsoever to believe 
that they took up their historical position so late. In their native tradition the 
Armenians trace their own existence in Armenia back until the eighth century 
BC, and references to them by their Iranian neighbours indicate an Armenian 
presence by at least the second century bc. The name Armenia (Arminiya) itself 
occurs in inscriptions as early as about 600 bc although one should always be 
wary of equating the name of a country with a people sharing that name. For the 
actual origins of the Armenian people we must rely on the evidence of their 
language, the historical testimony of their neighbours and predecessors, and 
archaeology. While all these approaches have not led to total consensus, there is 
broad agreement on the outline of the origins of the Armenians. 



Old Armenian 

Hayr mer or erkins: surb elici anun k'o. Ekesce ark'ayut'iwn k'o. 
Elicin kamk' k'o orpes erkins ew erki. 

Father our which in-heaven: holy become name thy. Come 
kingdom thy. 

Become will thy as in-heaven and on-earth .' 

20 The opening lines of the Lord's Prayer in Classical Armenian. The first word hayr is the 
Armenian reflex of Proto-Indo-European *p3ter where *p either changed to h, e.g., Proto-Indo- 
European *penk w e 'five* hut Armenian hing, or was usually lost, e.g., Sanskrit pad-, Greek 
pous, Latin pes and English foot but Armenian otn 'foot'. 

Although Armenian is clearly an Indo-European language, like Phrygian it 
shows no especially close relationship with the Anatolian languages other than 
borrowings. These are important since they indicate that proto-Armenians 
were in contact with both Luwian speakers and, more importantly, with 
Hittites. As Hittite was replaced by Luwian by 1200 bc, we may expect that 
proto-Armenians were passing through Anatolia before this time. The nearest 
linguistic neighbours to the Armenians were the Phrygians, Thracians and 
most especially the Greeks (some linguists even speak of Graeco-Armenian). 
All of this directs our attention towards the Balkans for their location prior to 
their migration through Anatolia. That the proto-Armenians were intrusive 
into their historical home is confirmed by the fact that they came to occupy the 
territory of the Urartians, the powerful kingdom of Van which flourished from 
the ninth to the sixth centuries bc and left abundant textual evidence that they 
were non-Indo-European speakers. Linguists today identify the Urartian 
language as closely related to Hurrian, the major non-Indo-European language 
of eastern Anatolia and northern Mesopotamia. Igor Diakonov has recently 
assembled evidence with indicates that the Armenians borrowed from the 
Hurro-Urartians words for slave, brick, seal, tin, local plant names (mint, 
pomegranate, plum, quince) and other items native to the region. The proto- 
Armenians apparently rose to power in the Armenian mountains after the 
collapse of the Urartian state and were certainly present there in the last half of 
the first millennium BC. The Armenian language then absorbed a vast quantity 
of foreign vocabulary from its neighbours, especially Iranian and Aramaic, the 
Semitic language spoken in north Mesopotamia, and the earliest of these loans 
appears to have taken place from the fifth to the first centuries bc. All of this 
suggests that the Armenians may have been part of the broad series of 
movements that also carried the Phrygians into Anatolia, and that by 500 bc 
they were establishing their pre-eminence in Armenia. 

In the most closely argued study of Armenian origins, the eminent Russian 
linguist and historian Igor Diakonov suggests that they may be identified with 
the Muski and perhaps other tribes known in Assyrian sources to have occupied 
the Upper Euphrates and lower Aratsani by 11 65 BC. These people appear to 
Diakonov to be in the right place at the right time, and their name offers tenuous 
links with both the Armenians themselves and with the Balkans. Muski is 



compared with the Georgian name for the Armenians, (Sa)mekhi, and with the 
ethnic formative found in both Mysia in Asia Minor and the Thracian province 
of xMoesia. Diakonov envisions that the Muski (proto-Armenians) drifted 
eastward absorbing both the Luwians of southern Anatolia and the Urartians in 
eastern Anatolia to form ultimately the Armenian people of today. 

The linkage of the Phrygians and Armenians to the same broad wave of 
migrations tangles their origins almost hopelessly. The Muski, identified by 
Diakonov in twelfth-century Assyrian texts as proto-Armenians, are just as 
probably assigned a Phrygian or Thracian identity by others, and without 
inscriptional evidence it is impossible to resolve the issue. It is useful to step 
back from the details and examine the broader patterns. 

It seems probable that a second wave of Indo-European migrations passed 
through Anatolia following that which produced the earlier Anatolian 
languages of Hittite and Luwian. Linguists do not find any cogent reasons for 
accepting an evolution of the earlier Hittite or Luwian languages into Phrygian 
or Armenian. Moreover, historical tradition such as that of Herodotus, coupled 
with the evidence of place names, for example, Thracian Moesia, Anatolian 
Mysia, and dialectal similarities with Greek, converge to suggest that these later 
Indo-European migrations derived from the Balkans. The evidence for these 
migrations in the archaeological record, other than the Balkan impact on 
northwest Anatolia, is not unequivocal, and we hardly need detail alternative 
explanations for the destruction horizons met in the twelfth century bc. But 
when Assyrian annals of the same century accuse the Muski of carrying out 
invasions with armies numbering 20,000 men, we have reason to accept the 
hypothesis that this was a period of migrating peoples. In two instances, we see 
clearly documented the linguistic replacement of one group by another - the 
Phrygians over the Hittites of central Anatolia, and the Armenians over the 
Urartians in eastern Anatolia. By the first centuries ad, all of these languages 
save Armenian have either long become extinct or are rapidly disappearing. 

The Indo-Aryans 

At first glance logic might seem to dictate that as we move east of Anatolia, the 
next Indo-European-speaking people we should encounter is the Iranians. It 

21 The concept of a common Indo- 
Iranian language is indicated by the 
close similarities between this Indie 
( Sanskrit) translation of an early 
Iranian hymn. The god M it raj 
Mithra was common to both Indians 
and Iranians. 

Avestan tarn amavantam yazatam 
Sanskrit tarn amavantam yajatam 

suram damohu savistam 
suram dhamasu savistham 

mithram yazai zaothrabyo 
mitram yajai hotrabhyah 

This powerful strong god Mithra 
strongest in the world of creatures, 
I will worship with libations. 



will soon become apparent that this is not so and that we are compelled to 
examine the evidence for Indian or Indo- Aryan origins first. In addition, when 
considering Indian origins we must keep in mind the broader ramifications of 
Indo-European dialectology. Probably the least-contested observation con- 
cerning the various Indo-European dialects is that those languages grouped 
together as Indie and Iranian show such remarkable similarities with one 
another that we can confidently posit a period of Indo-Iranian unity between 
the earlier Proto-Indo-European language and the subsequent appearance of 
the individual Indie (or Indo-Aryan) and Iranian languages. To these 
languages we may add the Kafiri languages of the Hindukush. Although these 
are only attested in recent times, they exhibit certain features that suggest that 
they are neither a direct descendant of Proto-Indo-Aryan or Proto-Iranian but 
an independant third branch of the Indo-Iranian group. 

Today Indie comprises the major languages of India and Pakistan, for 
example, Hindi-Urdu, Bengali, Panjabi, Marathi, and a host of others which 
total approximately 750 million speakers. These languages are essentially 
confined to the Indian subcontinent with the exception of more recent colonies 
that have expanded into Africa, the Pacific and Europe. The only exception to 
this is Romany, the language of the Gypsies, that was carried from northern 
India into Europe in the Middle Ages. 

The earliest written evidence preserved for Indo-Aryan in India only occurs 
about 300 bc with such monuments as the Asoka inscriptions. These, however, 
represent what linguists term Middle Indie, specifically Prakrit, and they by no 
means can serve as a terminus for the arrival of the Indians in the subcontinent. 
A vast literature of Old Indie, known as Sanskrit, preceded the Middle Indie 
inscriptions and formed the medium of the earliest Indie literary and religious 
language known to us. These were initially preserved only in oral form, but 

indrasya nu viryani pra vocam 
yani cakara prathamani vajri 
ahann ahim arm apas tatarda 
pra vaksana abhinat parvatanam. 

I shall proclaim now the heroic deeds of Indra, 
the first ones which the ctub-wielder performed, 
he slew the serpent, he made a breach for the waters, 
he split open the bellies of the mountains. 

2j A short excerpt from the Rig Veda celebrates the 
mar nor god Indra s victory over the evil Vrtra. The 
opening line contains a number of words with numerous 
cognates in the other Indo-European languages. For 
example, nu is the same as Greek, Old Irish, Lithuanian 
and Old English nu 'now'. Indra s heroic deeds viryani are 
'manly deeds' from the root vir- which is also found in 
Latin vir, Old Irish fer, Lithuanian vyras and Old English 
wer where it still survives in the compound 'werewolf. To 
proclaim, literally 'speak forth' pra vocam is cognate with 
Latin pro [forth 1 and voco 7 call'. 



there is abundant circumstantial evidence to indicate that they were written 
down by the sixth century bc. The earliest representative of this Old Indie is to 
be found in the Vedas, the ancient religious literature of India. The language of 
the V edas is very archaic, and the cultural and geographical world portrayed in 
these hymns suggests that they were composed in northwest India sometime 
before the first millennium bc with a notional date of around 1 500-1 200 BC. 

Although in India itself we can go no earlier than the Vedas, we have not 
thoroughly exhausted our sources, because the earliest written evidence for an 
Indo-Aryan language is not to be found in India but rather in northern Syria. 
Here lay the empire of the Mitanni who, by the fifteenth-fourteenth centuries 
BC, had expanded their power from the shores of the Mediterranean to the 
Zagros mountains thus coming into conflict with both the Hittites to the west 
and Egyptian control of the Euphrates. The language of the Mitanni was 
Hurrian, which we have already met in eastern Anatolia and north Syria. 
Although the basic language of the Mitanni was non-Indo-European, there is, 
nevertheless, clear evidence of the use of an Indo-European vocabulary in the 
Mitanni documents. These derive from diplomatic correspondence in foreign 
archives such as Bogazkoy (Hittite) and El Amarna (Egyptian), as the native 
Mitanni archives have not yet been discovered. But there can be little doubt that 
there was a distinctly Indo-Aryan element in the Mitanni kingdom. 

In a treaty between the Hittites and the Mitanni, the king of the latter swears 
by a series of Hurrian gods and then adds a series of names that are 
transparently the names of major Indie deities - Mi-it-ra (Indie Mitra), Aru-na 
(Varuna), In-da-ra (Indra) and Na-sa-at-tiya (Nasatya). A Hittite text on 
horse-training and chariotry, whose author is identified as Kikkuli the Mitanni, 
employs the names of Indie numerals for the courses that the chariot makes 
about a track- aika (Indie eka 'one'), tera (tri 'three'), panza {panca 'five'), satta 

24 The Mitanni and their 



'Thus (speaks) Kikkuli, the assussanni (horse-trainer), from the land 
of Mitanni: 

When he lets the horses onto the meadow in the autumn, he 
harnesses them. He lets them trot 3 miles, but he lets them gallop over 
7 fields. But on the way back he has them gallop over 1 0 fields. Then he 
unharnesses them, provides for them, and they are watered. He 
brings them into the stable. Then he gives them mixed together 1 
handful of wheat, 2 handfuls of barley and 1 handful of hay. They eat 
this up. As soon as they have finished their fodder, he binds them 
close to the post.' 

25 The opening instructions from the Hittite text on horse-tratmng by Kikkuli the Mitanni. In 
order to describe Kikkuli s profession, the text employs the Indo- Aryan word assussanni, 
(Sanskrit asvasani-^. 

(sapta 'seven') and na (nava 'nine'), while a Hurrian text from Yorgan Tepe 
employs Indo-Aryan words to describe the colour of horses, for example, babru 
(Indie babhru 'brown'), parita (palita 'grey') and pinkara (ptngala 'reddish'). 
The Mitanni word marya is precisely the same as the Vedic marya, 'warrior'. 

To these examples we may add a series of names for the Mitanni aristocracy 
and other names for divinities which associate the Indie element in the Mitanni 
language with the personal names and gods of the ruling dynasty. In addition, 
we naturally must include the special vocabulary of horsemanship for which the 
Mitanni were famous. These provide the clues that form the basis of the most 
widely accepted theory: an element of Indic-speaking chariot warriors 
superimposed themselves on a native Hurrian-speaking population to form a 
ruling dynasty that endured for several centuries. The precise mechanics by 
which this Indie element fused with the Hurrians to form the Mitanni is 
unknown, and scholars have employed everything from the model of an 
outright military conquest to the more benign analogy of the Varangian 
Norsemen who were 'invited' to establish the Old Russian state of Kiev. 

The date of the appearance of an Indie element in north Syria bears on any 
discussion of the origins and expansion of the Indo-Iranians in Asia. Although 
we possess texts from Mesopotamia in the preceding eighteenth and 
seventeenth centuries, there is no evidence in them that conclusively points to 
an Indie presence in the region. By the fifteenth century, we do have evidence 
for Indie elements in the Mitanni kingdom and there are also possible (though 
disputed) Indie traces in the names of a few gods revered by the Kassites, the 
dynasty from the Zagros region that assumed control of the Babylonian empire. 
By the thirteenth century the Mitanni kingdom collapses which sees an end to 
the Indie presence in Southwestern Asia. All of this suggests that the Indie 
element ascribable to the Mitanni (and perhaps to the Kassites) did not enter 
until about the sixteenth-fifteenth centuries bc. Furthermore, it attests the 
existence of a very archaic form of the Indie language as early as 1600 BC. This 
indicates that a separate Indo-Aryan language had already diverged from 
Iranian by this time and that the putative period of proto-Indo-Iranian 'unity' 
must predate this, perhaps by as much as a half millennium or more. This 
accords well with the broad estimates of historical linguists who believe that a 



continuum of Indo-Iranian languages probably began to diverge by 2000 bc, if 
not somewhat earlier. 

In tracing the origins of the Indie element in Western Asia we should begin 
with the obvious. The Indie presence is clearly not native but is intrusive to the 
Mitanni whose own language was Hurrian. The Kassites who apparently 
occupied the Zagros region before descending southwards into Babylonia also 
spoke a non-Indo-European language whose meagre textual remains do not 
permit clear association with any of the better-known languages of Asia. In 
either case, we are talking of intrusion, but one that is not so apparently 
thoroughgoing as among the Anatolians since it seems to have been limited to a 
small ruling dynasty which was otherwise assimilated by the native popula- 
tions. A solution to this nearly impossible archaeological puzzle was argued at 
length by the late Roman Ghirshman. 

Ghirshman observed that the area of Hurrian political power coincided with 
the distribution of a distinct ceramic type - Habur Ware. Within the territory of 
the Mitanni, and confined to the palaces of the aristocracy, there appeared this 
table ware, which included dark wares undecorated, impressed and white 
painted, none of which had a convincing local origin. Consequently, 
Ghirshman interpreted this pottery as the traditional ware of the Indie 
aristocracy in Mitanni and he found their closest parallels with wares from Shah 
Tepe in the Gorgan region southeast of the Caspian. We will see later that dark 
wares, specifically grey wares, have often been employed as an ethnic marker 
for other Indie and Iranian speakers. In addition, it is in the Gorgan region that 
the domestic horse first appears in the Near East about 3000-2250 b c. There is, 
for example, a cylinder seal depicting a horse-drawn vehicle from Hissar 1 1 IB. 
As the association of horsemanship and chariotry with Indie elements among 
the Mitanni has already been clearly established, this evidence could indicate 
the route of Indie movements towards Mesopotamia. Ghirshman concluded 
that at the end of the fourth millennium BC, a people carrying black wares and 
familiar with both the domestic horse and wheeled vehicles had penetrated 
northeastern Iran from the north. Here they subsequently developed chariotry 
and trumpets, a technological necessity of chariot warfare. They then gradually 
pushed south along the shores of the Caspian where they encountered the 
Hurrians about 1 800 BCatsuchsitesas Tepe Giyan 1 1 / 1 1 1 . This was the period 
of symbiosis of the Indie and Hurrian elements in the Zagros. They then 
expanded as part of the Hurrian movement into northern Mesopotamia where 
we have already alluded to their subsequent history. If one accepts the 
hypothesis, then by the third millennium bc the southern borders of either 
Proto-Indic or Proto-Indo-Iranian had already extended to the southeast 
Caspian and could be discerned throughout this region in the gradual adoption 
of grey wares. 

The primary ceramic evidence for migration of Indo-Aryans into the region 
of the Hurrians are the black and grey wares that appear abruptly in northern 
Mesopotamia and which Ghirshman derives ultimately from the southeast 
Caspian. This evidence has been challenged in two ways: either because grey 



26 Roman Ghirshman 
suggested that the Indo- Aryans 
migrated from southeast of the 
Caspian into the lands of the 

wares appear at Shah Tepe as early as 3000 bc, far too early to be assigned to the 
already differentiated Indo-Aryans; or else because grey wares in Iran itself 
expand too late and over too broad an area to be associated with anything other 
than the later Iranian migrations, not the earlier Indo-Aryans. Neither of these 
assertions, however, is wholly incompatible with Ghirshman's theory. It is 
theoretically possible that Proto-Indo-Iranian was spoken in the southeast 
Caspian about 3000 bc when grey ware first appears at Shah Tepe, but by the 
end of the next millennium, the language had gradually evolved into (Proto-) 
Indo-Aryan. The temporal crux of the argument is that migrations did begin 
when the urban centres of the Gorgan Grey Ware culture collapsed around 
1800 b c. The abandonment of major sites such as Tepe Hissar IIIC, Shah Tepe 
and Tureng Tepe was, according to Ghirshman, the result of nomadic pressure 
from the north which is reflected in the deposition of treasures on Gorgan sites 
and destruction horizons. The Gorgan refugees were forced around the 
Caspian and into the Zagros where they mingled with the Hurrians to emerge 
several centuries later as the Mitanni. Unfortunately, the ceramic evidence does 
not really support this migration, and the derivation of Mitanni dark wares 
from Gorgan Grey Ware remains distant and as yet unbridged by intermediary 
stages. Nor, of course, is there any prima-facie case for assuming that the 
expansion of the dark wares indicates a population movement much less a 
specifically Indo-Aryan movement. Moreover, the mechanics by which the 
Indo-Aryan element retained its ceramic forms but abandoned most other 
aspects of their language and culture to the Hurrians invites more scepticism 
than belief. 

Of Ghirshman's arguments, those pertaining to horse and chariot warfare 
demand our greatest attention, for here we find cultural elements that are 
inextricably associated with the specifically Indo-Aryan element in Mitanni 
and clearly set them apart from their non-Indo-European neighbours in the 
Near East. We have diplomatic correspondence between Egypt and Mitanni 
where the former requests both horses and chariots from the latter, indicating 



the Mitanni's reputation for horsemanship throughout the Near East. We have 
already seen in the linguistic evidence, such as Kikkuli's manual on 
horsemanship, that the terminology of chariotry included a distinctly Indo- 
Aryan vocabulary. Furthermore, the earliest evidence for the domestic horse is 
from the Pontic-Caspian region, and all present evidence suggests that it 
diffused from there through the Caucasus into Anatolia and perhaps around the 
eastern Caspian into northeast Iran. Hence, the appearance of both the horse 
and the chariot have frequently been attributed to an expansion of Indo- 
Europeans from the north into Western Asia. Specifically, Indo-Aryan 
charioteers are seen penetrating the kingdoms of Southwest Asia where, in the 
case of the Mitanni, they were able to dominate the local Hurrian substrate. 
Naturally, the evidence is anything but so certain. 

We know that wheeled vehicles were employed in Mesopotamia by 3000 bc 
in early Sumer, and their presence in southern Mesopotamia has no obvious 
direct association with the Indo-Europeans. These vehicles were basically 
drawn by bovids, although there was a gradual increase in the use of equid 
draught in Western Asia. This, however, was primarily the onager or ass, and at 
no time prior to the second millennium bc can we regard Southwest Asia as 
practising the horse- and chariot-centred warfare that one finds among the 
Indo-Aryans. The earliest evidence for the horse in Western Asia is presently 
limited to Tal-i Iblis in south-central Iran (3500 bc) and Selenkahiyeh in Syria 
(2400-2000 bc), and its attestation in cuneiform texts appears to be similarly 
late and dates to the end of the third millennium bc. But from early in the 
second millennium bc we find unequivocal evidence for both the horse and the 
chariot, and by the seventeenth-sixteenth centuries this form of warfare is 
found from northern Anatolia south to Nubia, which illustrates the rapid 
spread of this revolutionary technology. J.H. Crouwel and M. A. Littauer have 
argued that this evidence suggests a perfectly logical evolution of the two- 
wheeled cart into the spoked-wheel chariot within Western Asia itself prior to 
the appearance of the Indo-Aryans whose presence in this region cannot be 
demonstrated before about 1600 BC. Some scholars such as Diakonov go on to 
argue that there is consequently no case for employing the earliest appearance 
of the domestic horse and chariot in the Near East as an ethnic marker for Indo- 
European migrations. 

No easy solution presents itself since the problems here involve at least three 
issues which need not necessarily be linked - the origin of the domestic horse in 
Southwest Asia, the origins of the chariot, and the date of Indo-Aryan 
movements into the region. Current evidence provides little reason for seeing 
the domestic horse as anything but intrusive into Southwest Asia. The paucity 
of sites on which horse bones appear is in sharp contrast to the abundant 
evidence of the domestic horse during the fourth-third millennium across the 
Pontic-Caspian-Siberian steppe, and the most economical argument is to 
derive domestic horse populations from this direction during the third and 
early second millennia. The chariot, on the other hand, could possibly have 
been invented independently in the Near East as well as in the steppe. There is, 



I think, much to be said for Stuart Piggott's suggestion that the Near Eastern 
war-chariot may have been the result of a symbiosis of the local needs and 
tradition of battle vehicles brought into contact with the dispersal of horse- 
drawn vehicles employing spoked wheels typical of the steppe. In any event, the 
horse-drawn war-chariot need not be regarded exclusively as an Indo- 
European ethnic marker, especially as it spread so rapidly over a vast area. But 
we should not be too quick to exclude the possibility that the earliest chariots 
were associated with Indo- Aryans because the Indo-Aryans are not attested in 
the Near East until several centuries after the appearance of chariots. Our 
dating of the Indo-Aryan element in the Mitanni texts is based purely and 
simply on written documents offering datable contexts. While we cannot with 
certainty push these dates back prior to the fifteenth century bc, it should not be 
forgotten that the Indie elements seem to be little more than the residue of a 
dead language in Hurrian, and that the symbiosis that produced the Mitanni 
may have taken place centuries earlier. On such an issue, where the discovery of 
a single datable text could advance the Indo-Aryans to greater antiquity in the 
Near East, it is not wise to rush to judgment on the issue of chariot warfare. We 
will return to the Mitanni problem when we have expanded our survey of the 

Before we turn to the Indian subcontinent, we must briefly tackle the 
obvious problem of the relationship with the Indo-Aryans of Western Asia and 
those of India itself. Of three possibilities, received opinion rejects two of them. 
It is highly improbable that the Indo-Aryans of Western Asia migrated 
eastw ards, for example with the collapse of the Mitanni, and wandered into 
India, since there is not a shred of evidence - for example, names of non-Indic 
deities, personal names, loan words - that the Indo-Aryans of India ever had 
any contacts with their west Asian neighbours. The reverse possibility, that the 
main line of migration was into India and that a small group broke off and 
wandered from India into Western Asia is readily dismissed as an improbably 
long migration, again without the least bit of evidence. Having excluded the 
unlikely, we are left with the merely possible - either the Indo-Aryans divided 
south of an earlier staging area with some moving east and others far to the west, 
or they actually immigrated in mass forming a broad continuum across Western 
Asia to the Indus and were later divided by the incursion of Iranian-speaking 
peoples. The first explanation is still along those lines advanced by Ghirshman 
and others while the Indologist, Thomas Burrow, has argued for an initial 
Indo-Aryan settlement not only of north Mesopotamia but also of the Iranian 
plateau itself. The arguments are primarily linguistic and religious, among 
which the latter is the most intriguing. 

Burrow sets out to explain why the Indie word for god, deva y should occur 
under its Iranian cognate daeva to mean 'demon'. The Iranian situation was the 
result of the great religious reformer Zarathustra's (or Zoroaster as he is known 
through Greek sources) influence on the Iranian religion which, according to 
Burrow, resulted in the relegation of those specifically Indo-Aryan gods such as 
Varuna and Indra to the role of demons, since they were the gods of the 



indigenous Indo-Aryan occupiers of Iran and were not recognized as gods in 
the new Iranian religion. This argument goes much further, because 
Zoroastrian religion employs a special class of words to describe demons. Some 
of these words are clearly formed pejoratives, but a small set of them appear to 
derive from an earlier Indo-Aryan substrate. Further evidence is found in the 
names of Iranian rivers which are seen to be borrowings from an earlier Indo- 
Aryan language and attested in India itself. In short, Zoroastrianism was a 
religious-based crusade against the remnant Indo-Aryan population which 
occupied Iran, and it was the success of this Iranian expansion that split the 
Indo-Aryans into western (Mitanni) and eastern (Indian) groups. 

Burrow argues that the conventional dating of Zarathustra to about 600 bc is 
far too recent and that there are reasons for placing him half a millennium 
earlier at least. Hence, Burrow would argue that the Indo-Aryan continuum 
that spanned Southwest Asia and Iran was established before 1400 bc by which 
time the Iranians were already beginning to expand into northeastern Iran. 

While a number of the specifics of Burrow's theory have already been 
dismissed by some linguists, his suggestion that there must have been an Indo- 
Aryan element in the substrate later subsumed by the expanding Iranians is still 
attractive. Unfortunately, the theory has been proposed almost exclusively on 
linguistic grounds and there has been little attempt to seek archaeological 
correlations for his scheme. This is not at all the problem, however, when we 
turn our attention to the intrusion of Indo-Aryan speakers into India itself. 

As we have done with the other Indo-European peoples of Asia, we must 
preface our inquiry into the origins of the Indo-Aryans in India with the 
problem of their autochthon) . Although the great majority of scholars insist 
that the Indo-Aryans were intrusive into northwest India, there have always 
been a few to claim that the Indus Valley civilization that flourished about 
2500-1500 bc was Indo-Aryan and it must be admitted that direct written 

27 The area of the Indus 
civilization ( Jotted line) falls 
within the general distribution 
of the ELuno-Dravidian 
language family. Mote that the 
Oravidian languages ate not 
only found to the south of the 
Indus hut Brahui a northern 
Dru vidian language, lies to 
its west. 

Brahui Dravidian Languages 



testimony is not so clear in India as it is in much of Southwest Asia. This is 
because the only written evidence left by the builders of the Indus towns - the 
Indus script - remains undeciphered despite periodic claims to the contrary. It 
is possible that the brevity of Indus texts, largely confined to seals, coupled with 
the absence of bilingual inscriptions will ensure our ignorance indefinitely. 
Nevertheless, linguists also recognize that there are some linguistic solutions to 
the problem of the Indus script that are far more likely than others. 

It should be remembered that as well as Indo-Aryan languages, India also 
possesses two other major linguistic groups. By far the largest are the non-Indo- 
European Dravidian languages which dominate the southern third of India. 
The positioning of Indo-Aryan in the north, and the known historical 
expansion of Indie from north to both the east and the south, gives us every 
reason to deny the Indo-Aryans a prior home in those regions. Furthermore, 
there are still remnant northern Dravidian languages including Brahui to the 
west of the Indus and others to the southeast. The most obvious explanation of 
this situation is that the Dravidian languages once occupied nearly all of the 
Indian subcontinent and it is the intrusion of Indo-Aryans that engulfed them 
in northern India leaving but a few isolated enclaves. This is further supported 
by the fact that Dravidian loan words begin to appear in Sanskrit literature from 
its very beginning. This prior dominance of Dravidian in northern India makes 
it the most formidable candidate for the language of the Indus civilization. 
Much less likely is Munda, a non-Indo-European language now confined to 
central India but which once extended north to the Ganges. It appears to have 
been itself swamped by Dravidian languages and then further confined by the 
expansion of the Indo-Aryans. 

Another candidate at least proximate to the Indus Valley is Elamite, the 
language of the major kingdom that occupied southern Iran. Here we are 
dealing with a literate society writing in a decipherable non-Indo-European 
language (related to Dravidian) which has left us intelligible texts from the late 
third millennium onwards. Within a number of Elamite sites there also occur 
pictographic tablets, related generically to the earliest Sumerian script, and 
dating as early as the late fourth millennium B c. They have been found on sites 
as far east as Tepe Yahya and Shahr-i Sokhta. These tablets are generally 
termed Proto-Elamite since they coincide with the location of later decipher- 
able Elamite texts. For those requiring an exterior high culture to explain the 
origins of the Indus civilization (generally regarded today as unnecessary), 
Proto-Elamite can always serve as a popular candidate which offers yet another 
proximate linguistic identity for the Indus script. Whatever the merits of these 
arguments are, the existence of Elamite across southern Iran excludes this 
region as an earlier homeland of the Indo-Aryans. 

Circumstantial evidence for identifying the language of the Indus Valley 
script with Elamite or Dravidian has been greatly strengthened by David 
McAlpin's work on the relationship between the Dravidian languages and 
Elamite. McAlpin has demonstrated that the two groups of languages derive 
from a common proto-language, Proto-Elamo-Dravidian, and that Brahui, 



traditionally assigned to the Northern Dravidian subgroup, would actually 
appear to be linguistically as well as geographically intermediate between the 
two major subgroups, McAlpin reconstructs to Proto-Elamo-Dravidian a 
common stockbreeding vocabulary (cattle, ovicaprid, goat, and so on), and 
traces special developments in the agricultural terminology of the Dravidian 
branch as it pushed southward into the Indian subcontinent. It is quite 
interesting that one may reconstruct a common word for 'brick' in Proto- 
Elamo-Dravidian as this was the ubiquitous building material of the Indus and 
other neighbouring civilizations. McAlpin dates the disintegration of Proto- 
Elamo-Dravidian to about the fifth millennium bc. All of this makes a good case 
for associating the early village farming economies that formed the foundation 
of the Indus civilization with Elamo-Dravidian languages - an hypothesis far 
more probable than Colin Renfrew's recent suggestion that the Indus Valley 
civilization was Indo-Aryan and that it was Indo-Europeans who introduced 
the farming economy to this region. 11 

Other than the Indus script, the very character of Indian society reflected in 
the earliest Vedic literature renders it highly unlikely that the Indus civilization 
was the product of Indo-Aryans. Although the earliest Vedic hymns are 
focused geographically on the Indus and its major tributaries in the Punjab, the 
culture represented in them bears little similarity to that of the urban society 
found at Harappa or Mohenjo-daro. It is illiterate, non-urban, non-maritime, 
basically uninterested in exchange other than that involving cattle, and lacking 
in any forms of political complexity beyond that of a king whose primary 
function seems to be concerned with warfare and ritual. Moreover, the Vedas 
recognize a dichotomy between the Indo-Aryans and their dark-skinned 
enemies, the dasa, who are on one occasion described as 'nose-less', which has 
generally been interpreted as a pejorative reference to Dravidian physical 
features. Vedic hymns commemorate or invoke divine support for the 
destruction of their enemies and the storming of their citadels. This is to be 
accomplished with the assistance of their horses and chariots, a technique of 
warfare apparently unknown to the Indus civilization. It is little surprising, 
then, that the Indo-Aryans have been viewed in the past as the probable 
destroyers of the Indus civilization about 1 500 b c. Today, there is a tendency to 
seek internal causes for this collapse - climatic, hydraulic, environmental - 
rather than Indie warriors who are more often seen as a phenomenon of the 
post-Indus period. In either case, the only way that one may retain an Indo- 
Aryan identity for the Indus civilization is to assume that, after its collapse 
about 1500 bc, it receded into the type of world reflected in the Vedic hymns 
and that these are the product of the degenerate descendants of the Indus 
civilization. Given all the other objections, this solution would call for far more 
special pleading than anyone has reason to credit. All of our earliest evidence for 
the Indo-Aryans in India, therefore, indicates that they came from elsewhere 
and we should turn to the archaeological evidence to trace their migration. 

A reasonable starting point is to assume that whatever culture occupied the 
lands depicted in early Indie literature in the first millennium bc has the most 



right to the label oflndo-Aryan. The Painted Grey Ware culture, centred from 
the eastern Punjab to the central Ganges, is at least one obvious candidate. The 
radiocarbon dates indicate that this culture flourished about 1300-400 bc, a 
time when we can be certain of the existence of Indie speakers in northern India 
without fear of over-extrapolating from the literary evidence. The culture, 
which takes its name from its fine grey ware painted with black or red 
decoration, meets some of the minimum requirements of an Indo-Aryan 
culture as seen through the earliest Indie literature. Settlements, where 
attested, tend to be of flimsy wattle and daub and bear no resemblance to the 
brick-built urban complexes known in the Harappan culture. The economy 
included the domestic horse, and although this animal has occasionally been 
recovered from Harappan sites, for example Surkotada and Kalibangan, no one 
would credit the earlier Harappan culture as exemplifying the horse-centred 
culture of the Vedic Aryans. Furthermore, there is an excellent correlation 
between the earliest Painted Grey Ware sites and historical sites mentioned in 
the great Indie epic, the Mahabharata, which according to tradition is set to an 
historical period of the early first millennium bc. That these sites also include 
what would later become major urban centres coincides w ell with their identity 
as early Indo-Aryan settlements which rapidly spread their language in 
northern India. The major problem with identifying them with the earliest 
intrusive Indo-Aryans is that the ceramics are without clear external as well as 
internal derivation and so it is difficult to postulate an invading culture. 
Furthermore, the remains largely date to the first millennium bc and are 
primarily concentrated to the east of where w e w ould expect to find the earliest 
Indo-Aryan remains in India. For this we must look earlier and further west. 
There are just such candidates. 

28, 2g The Gandhara Grave, Cemetery H and Painted Grey Ware cultures are traditional 
candidates for Indo-Aryan (or Kafiri) origins; seen here is a Gandhara Grave burial. 



The Cemetery H culture of the middle Indus has been advanced as the 
possible archaeological manifestation of the Indo-Aryan invaders. Its temporal 
position is not entirely secure other than being post Harappan, and certainly 
some of the stratigraphic evidence of the Cemetery H remains suggests a 
significant interval between the fall of Harappa and this culture. This is not in 
itself a hindrance, since we have already seen that there really is no compelling 
reason to credit the Indo-Aryans with the destruction of the Indus civilization, 
although they have played out that role in a number of earlier archaeological 
models. It is their pottery with possible West Asiatic connections coupled with 
their fractional burial rite, that is, the collection of partial skeletal remains 
following exposure and their deposition in an urn, which has cast them as 
intruders. In addition, what little architectural remains exist also support the 
image of a less sedentary population than their Harappan predecessors. But 
their limited distribution and insecure foreign affinities do not encourage us to 
accept them as the whole solution to the Indo-Aryan problem. 

One of the best candidates for intruding Indo-Aryans is to be found north of 
the Indus in the Swat Valley, which gives its name to the culture otherwise 
known as the Gandhara Grave culture. The Swat Valley occupies the position 
of a trip line, in that any intruder from the northwest is likely to pass through it 
first before arriving in either the Indus or Ganges Basin. Consequently, a major 
cultural change here at the appropriate time might signal the arrival of the Indo- 
Aryans. Just such a cultural break is argued to occur about 1800 BC with the 
introduction of a new burial rite and ceramics into the Swat Valley. The culture 
is known from cemeteries where we find both flexed inhumation in a pit 
(actually the sub-pit of an initial pit) and, more interestingly, cremation burial 
in an urn, often a face-urn. Such a rite attracts our attention not only because it 
is not found in contemporary cultures of the same region but also because early 
Vedic literature indicates that both inhumation and cremation burial were 
practised in early Indo-Aryan society. The goods found with the burials are not 
especially rich but do include copper, gold and silver in the earlier period, and 
iron is found by about 900 b c. The settlements indicate semi-subterranean huts 
in the initial phases with the later introduction of rubble-filled walls. The 
economy included a range of domestic plants and animals, among the latter the 
horse. It should be emphasized that the horse was not merely incidental to the 
faunal remains but we also have two horse burials as well as horse-trappings 
from the Gandhara Grave culture. Furthermore, the new ceramic style was a 
grey ware, approximately half of which was handmade and decorated with 
incisions. These ceramics show a similarity to grey wares of south Central Asia 
and northern Iran, precisely the direction from which we might expect the 
earliest Indo-Aryans. Moreover, the Swat region maintains its cultural 
continuity down to about 400 BC and, consequently, the Swat Valley culture 
offers itself as a most attractive candidate for early Indo-Aryans or Kafiri if we 
are correct in assuming that they must have migrated through this region. 
Finally, the area makes an excellent fit with the geographical scene depicted in 
the hymns of the Rig Veda and it does so at the expected time. 




Naturally, our optimism at having identified a suitable archaeological 
expression of the Indo-Aryans must be tempered with some caution. It has not 
yet been possible to make a convincing association between the Swat Vallev 
culture and any of the putative Indo-Aryan cultures that appear later in the 
Punjab or the Ganges Basin. The leap from the plain grey wares of the Swat 
Valley to the Painted Grey Ware culture of India, suggested by the Pakistani 
archaeologist Ahmad Hassan Dani, has not been accepted by Indian 
archaeologists. Moreover, the connection between the Swat Valley and the 
northern manifestations of the grey ware tradition is also in need of 
strengthening. We should also recall that it is out of this grey ware tradition that 
many also derive the Iranians, and it is specifically to their origins that we 
should now turn. 

The Iranians 

The Iranian languages, for example, Persian, Kurdish and Pashto, are today 
primarily confined to the modern states of Iran and Afghanistan, and to 
territories immediately adjacent, all of which comprise over seventy million 
speakers. The current distribution of Iranian, however, greatly belies its earlier 
expansion which included a vast portion of the Eurasian steppe. Reading from 
west to east we can include as Iranian speakers the major Iron Age nomads of 
the Pontic-Caspian steppe such as the Kimmerians (?), Scythians, Sarmatians 
and Alans. The incredible mobility of these horse-mounted nomads becomes 
all the more impressive when we recall their westward expansions through 
Europe. Sarmatian tribes not only settled in the Danube region but, during the 
second century ad, were conscripted to defend the borders of Roman Britain. 
The Alans travelled as far west as France and forced their way south through 
Spain, ultimately to establish a state in North Africa. Of these different peoples, 
only remnants of the Alans have survived to the present day in the modern guise 


of the Ossetes who, retreating into the protection of the central Caucasus in the 
Middle Ages, still retain a population of nearly 600,000 speakers. Much further 
to the east were other Iranian-speaking peoples, such as those who have left us 
the remains of Sogdian, language of the ancient kingdom of Samarkand, first 
attested in the early Middle Ages and which still survives among the Yaghnobis 
of Tadzhikistan. Furthest to the east lay the bearers of Khotanese Saka, an 
Iranian language situated in Chinese Turkestan which was employed in 
religious texts of the seventh to tenth centuries and is preserved today among 
the Pamirs. This extremely broad group of steppe languages are all labelled 
Northeastern and Eastern Iranian by linguists, to contrast them with the more 
familiar Western Iranian language of Persian. 

The earliest Western Iranian speakers emerge into history in northwestern 
Iran in the mid-ninth century bc. This is when Assyrian texts record the 
existence of the Medes (836 bc) in the vicinity of modern Isfahan and the 
Persians (Parsua) northwest of Kermanshah. At this time the twenty-seven 
kings of the Parsua are recorded delivering tribute to the Assyrian 
Shalmanesser III (858-824 bc). It is clear from the Assyrian narratives of both 
the Medes and the Persians that they are encountering already established 
peoples with whom they came into contact and conflict as the Assyrians pushed 
northeastwards to the Zagros. We may assume, therefore, that Iranian tribes 
were already settled by the beginning of the ninth century in the region north of 
the Zagros, and it is here that we can expect to pick up their trail. 

The period in which we first encounter the Iranians in western Iran is 
designated Iron Age Period II (1000-800 bc). This in turn is seen as the direct 
and uninterrupted successor to Iron Age I which begins about 1400 bc. It is the 
initiation of Iron Age I that sees a major cultural break in this region. With its 
earliest appearance we find a shift from painted to plain grey wares, both in 
settlements and as grave accompaniments. The cemeteries themselves, such as 
Cemetery A at Sialk (V), mark a change from the intramural burials earlier 
encountered in the area. Iron is almost non-existent in the earliest phase but 

Auramazda vazraka hya mathista baganam 

hauv Darayavaum XSyam 

ada hausaiy xsacam 

frabara tya naibam 

tya uratharam uvaspam umartiyam 

Great Ahuramazda, the greatest of gods - 

he created Darius the King, 

he bestowed upon him the kingdom, 


possessed of good charioteers, of good horses, of good men. 

31 A brief insc ription of Darius in the Western Iranian Old Persian language. The last line 
contains three words with the prefix u ' good 1 followed by ratha "chariot 1 (cf Sanskrit ratha- 
Latin rota, Irish roth, Lithuanian ratas 'wheeF; aspa 'horse ' (Sanskrit asva-, Latin equus, Old 
Irish ech, Lithuanian asva,/; and martiya 'man (Sanskrit marta-, Greek mortos, from the root 
*mer- 'to die, cf mortal). 



becomes increasingly abundant by Period II. Most importantly, the area in 
which the grey ware is found coincides in general with the later Assyrian 
evidence for Iranian tribes. On the basis of this evidence, T. Cuyler Young 
concluded that the Iron Age I culture north of the Zagros represents a sharp 
cultural break w hich should be associated with the emergence of the Iranians in 
western Iran. The continuity of this culture through Iron Age II sees trends 
towards greater regionalization and the absorption of foreign influences, which 
coincides well with the Iranians' contact with the Assyrian world. In short, 
some archaeologists argue that we have a relatively good association between an 
apparently intrusive culture and the historic distribution of an Indo-European 
people. It is when we try to trace the origins of the Western Iranians further 
back in time that our problems become much more difficult. 

Since the Iranians are first encountered north or east of a chain of non-Indo- 
European Urartians, Assyrians, and Elamites, their approach most probably 
was from the north. We can readily exclude the northwest, that is, eastern 
Anatolia, as highly improbable given its prior identification with Hurrian 
populations. As we have seen, the possibility that the Western Iranians derived 
ultimately from the Pontic steppe was entertained by Ghirshman and others 
who argued for a migration across the Caucasus and down the western Caspian 
into northern Iran. The evidence offered to support such a migration is varied. 
There are linguists who note the occasional presence of Iranian loan words 
among the Caucasian languages along the proposed route. For their part, some 
archaeologists cite the appearance of steppe kurgans, dating to the centuries 
around iooo bc, that appear in northern Azerbaijan. These, with their wooden 
burial structure, hearths and horse burials, are similar to those found in the 
Srubnaya (Timber-grave) culture north of the Caucasus as well as some burials 
in northwestern Iran. Moreover, the proposed migration route is precisely that 
which is historically attested for Iranian-speaking Scythians who penetrated 
from north of the Black Sea through the Caucasus and on into Southwest Asia. 
Nevertheless, many are still very sceptical that any secure link can be 

Haftavan Ya nik ~ L 
Dinkha . Hasanlu /v * 


: . . Yarim Tepe 

^ Marlir%o,:<; Shah Tepe • 
Dinkha •Hasanlu ^ ^g^^ TurengTepe 

Medes ^^!^ A . TepeHissar 

0 300 

v ivieaes -y* 
^•Godin V ' 

TepeGiyan ^sialk 


32 Grey-Ware sites are frequently 
associated with West Iranian 
migrations, possibly originating from 
the southeast (Caspian. Xote that 
the distribution of these sites exceeds 
the area of historically attested 



established between the Iron Age I culture of Iran and either the Caucasus or 
Pontic steppe, and they would dismiss outright any claim for a full-scale 
migration along this route about iooo bc. This leaves but one other source - the 
northeast, a route which Diakonov argues is the most plausible since it involves 
passage between similar ecological regions (Central Asia and Iran), without 
major barriers. 

We have already seen how the grey ware tradition had appeared in the 
Gorgan region before the end of the fourth millennium, and how it seems to 
terminate about 1800 BC (Hissar IIIC) coincidental with the abandonment of 
Bronze Age sites in the southeast Caspian. We have also reviewed how grey 
wares again begin to emerge about four centuries later in Iron Age I on sites 
further to the southwest, which coincides with the territory we would assign to 
the earliest historic Iranians. The most obvious and economical solution then 
would be to assume a progression - by migration or diffusion - of these wares 
from the Gorgan towards the Zagros. Some archaeologists, perhaps crediting 
radiocarbon dating with far greater precision than the technique usually 
delivers, have been troubled by the apparent 400-year hiatus between the final 
Gorgan material and earliest Iron Age I dates. Even if this hiatus is justified on 
the present evidence, some believe that it will be bridged in the course of future 
excavations. In any event, it is to the northeast that we naturally would seek the 
Iranians, since it is in this area that we subsequently find other major Iranian 
peoples such as the Parthians, Bactrians and Sogdians. Moreover, Diakonov 
argues that the absence of foreign words in the earliest Eastern Iranian language 
indicates its longer occupation of the area than Western Iranian which abounds 
in words drawn from its non-Iranian neighbours and substrates. Attractive as 
this solution might appear, we must admit that it does contain one major 
drawback - it sets the Iranians immediately on the heels of the Indo- Aryans. 

We have seen how Ghirshman and others provided the Western Iranians 
with an origin in the Pontic-Caspian steppe while deriving the Indo-Aryans 
from Central Asia. Although we may be sceptical of the archaeological evi- 

kam namoi zam 
pairi x v aetaus 
noit ma xsnaus 
naeda dahyaus 
katha thwa 

yoi sastaro dragvanto 
mazda xsnaosai ahura 
ahmi mazda anaeso 
hyatca kamnana ahmi 

kuthra namoi ayeni 

airyamanasca dadaiti 
ya varazana haca 

vaeda tat ya 
ma kamnafsva 

33 An excerpt from 
Zarathustras Gathas, the 
oldest part of the Avesta. 

To what land shall I flee? Where bend my steps? 

I am thrust out from family and tribe; 

I have no favour from the village to which I belong, 

Nor from the wicked rulers of the country: 

How then, O Lord, shall I obtain thy favour? 

I know, O Wise One, why I am powerless: 

My cattle are few, and I have few men. 



dence adduced to demonstrate an Iranian migration across the Caucasus in the 
second millennium bc, it must be admitted that such a solution at least 
accommodates the linguistic divergence undergone by the Indo-Iranian 
languages. Derivation of both the Western Iranians and the Proto-Indo- Aryans 
from the Grey Ware tradition of the southeast Caspian offers little room for 
geographical separation between the two branches of Indo-Iranian. Only by 
detaching the Indo-Aryan element in Mitanni from a possible origin in the 
Gorgan culture can we avoid the linguistic implausibility of deriving both 
Indians and Iranians from precisely the same region at the same time. Or, of 
course, we can accept the Grey Ware identification with the Indo-Aryans and 
seek a different origin for the Iranians. For this, we must consider the origin of 
the Eastern Iranians. 

It is unfortunate for us that the earliest native literature in an Iranian 
language is situated so far from those historical sources of the Near East that 
recount the earliest appearance of the Western Iranians. Avestan, the first 
attested Iranian language, takes its name from the great body of early Iranian 
religious literature. The earliest portion of the Avesta is the Gathas, which are 
credited to Zarathustra. This collection of hymns, localized geographically to 
Central Asia/northeastern Iran, displays much the same archaic nature as do 
the Vedas. It is this which has prompted linguists such as Burrow to reject the 
traditional dating of Zarathustra to the sixth century bc and to propose a much 
older date, possibly half a millennium or more. The nature of the dating 
controversy provides little room for compromise between the opposing forces, 
and from the archaeological perspective we can at least credit the earliest part of 
the Avesta to the Late Bronze Age-Early Iron Age, about iooo bc. 

The Gathas are fundamentally religious hymns and do not provide as rich a 
source for ethnographic inference as do either the later portions of the Avesta or 
the Rig Veda. What can be gleaned from the hymns is that they appear to be 
composed in an essentially rural milieu where stockbreeding, especially cattle- 
keeping, is held in esteem along with agriculture. Urbanism of any sort is not 
suggested although mention of citadels, irrigation canals, and the like, are met 
in other later Avestan hymns. The hints of political structure represented in the 
hymns are limited to family, village and tribe, or to district and country. Finally, 
it is commonly maintained that the Gathas reflect an opposition between the 
believer in Zarathustra's teachings and unbelievers, not between Iranians and 

If we search for an archaeological correlate for the world of the Gathas, then 
the area assigned by the few geographical references in the text fits well enough. 
By the mid-second millennium BC, most of the northern regions of Central Asia 
were occupied by pastoral societies deriving, either from the Timber-grave 
culture which was centred west on the Volga, or the Andronovo culture, a 
blanket term for a variety of steppe Bronze Age cultures that emerged in Central 
Asia and south Siberia. What is important for our purpose is that they are in 
sharp contrast to the proto-urban centres of southern Central Asia such as 
Namazga V or Altin Depe, whose sophisticated architecture including temples, 


technology, art and economy bear scant resemblance to that portrayed in the 
earliest Iranian literature. Moreover, the remains from these steppe Bronze Age 
sites provide us with some of the finest parallels with common reconstructions 
for Indo-Iranian culture. The settlement and cemetery of Sintashta, for 
example, although located far to the north on the Trans-Ural steppe, provides 
the type of Indo-Iranian archaeological evidence that would more than delight 
an archaeologist seeking their remains in Iran or India. Next to a small 
settlement occurs a cemetery of tumulus burials dating to the sixteenth century 
bc. These contain the remains of large quantities of sacrificed animals, 
especially horses and dogs which are noted in Indo-Aryan ritual, evidence of 
chariots, and an assortment of other Indo-Iranian ritual markers. 

The identification of the Andronovo culture and at least the eastern outliers 
of the Timber-grave culture as Indo-Iranian is commonly accepted by scholars. 
It is out of this steppe region that we derive the Scythians who pushed 
westwards to the Pontic and then southwards through the Caucasus to ravage 
the Near East during the seventh century bc. This migration is traceable both 
in historical sources and archaeological remains such as burials and Scythian 

To the east of the Scythians emerged the Sarmatians, Massagetae, and Alans, 
all of whom can be derived from the Late Bronze and Early Iron Age of the 
Asiatic steppe. Parthians, Bactrians, Sogdians and other Iranian peoples known 
to us from Central Asia emerge out of the Iron Age cultures of their respective 
regions which seem to have involved the acculturation or evolution of earlier 
pastoralists to the increasingly urbanized or citadel-based settlements that 
emerged in the region in the first millennium bc. As the eastern limits of the 
Andronovo culture extended all the way to the Yenisey river, Kirghiziya and 
Tadzhikistan, one might even be able to provide a not too distant source for the 
easternmost Iranians, the Saka. Indeed, it is in the eastern Andronovo variants 
such as the Bishkent culture of south Tadzhikistan that one encounters again 
the probable expression of Indo-Iranian ritual in the archaeological record. At 
the cemetery of Tulkhar, male burials were provided with small rectangular 
hearths, reminiscent of the typical Ahavaniya, the rectangular fire-altar of early 
Indie priests, while females were provided round hearths, comparable to the 
Garhapatya, the female-associated hearth fire of the Indo-Aryan house. 

Even if there is considerable disagreement as to detail, the identification of 
the steppe Bronze Age cultures as essentially Indo-Iranian seems fairly secure. 
This security, however, only remains if we are deliberately vague and do not 
specify whether we mean Indo-Iranian, Indo-Aryan or Iranian. It is often 
reasoned that since the steppe cultures date to the second millennium or later, it 
is most probable that they are (Proto-) Iranian rather than undifferentiated 
Indo-Iranians. Justification for this derives from our knowledge that Indo- 
Aryan had already emerged by the second millennium bc among the Mitanni. 
Having now come full circle, we must step outside the more specific arguments 
and briefly view this maze into w hich both linguists and archaeologists have got 



34, 35 A male and a female burial from Tulkhar. The male was buried with a rectangular 
hearth and with bones of a sheep, a dagger, pot, bead-amulet and a flint arrowhead; the female 
was buried with a round hearth, sheep bones and a pot. 

All archaeological evidence adduced to indicate Indo-Iranian migrations 
derives broadly from two different sources - grey wares associated with proto- 
urban settlements extending from south of the Caspian across the southern 
border of Central Asia, and traces (often burials) of either mobile pastoralists or 
smaller villages, be they in the Eurasian steppe or in northwestern India. The 
evidence of the Avesta and the Vedas would clearly point to the second category 
of remains as the more probable archaeological expression of Indo-Iranian 
culture. Consequently, a number of Soviet scholars such as Edvin Grantovsky 
have rejected outright any ascription of grey ware cultures or any other 
evidence derived from the first category of sites to the Indo-Iranians as 
incompatible with their pastoral origins. But we should not forget that our 
cultural reconstructions are based primarily on an East Iranian text, the Avesta, 
and the Vedas, both of which were composed in regions remote from either the 
Indo-Aryans or western Iranians in the Near East. Here our literary evidence is 
very meagre. In short, we have no conclusive grounds to deny the Western 
Iranians or the Indo-Aryans of the Near East either an immediate urban past or 
at least a close association with urban societies. We cannot escape seriously 
entertaining all of those arguments that seek to trace population movements by 
the spread of essentially Central Asian ceramic styles, especially when 
alternative evidence deriving from pastoralists is as yet so poorly known in those 
areas of the Near East later occupied by Indie or Iranian speakers. 

The use of grey wares to indicate migrations in Iran can, as we have seen, lead 
to apparent contradictions. It would seem improbable that the Indo-Aryans of 
Mitanni moved west with the same expansion of grey wares as the Western 
Iranians. Any increase in the antiquity of grey wares north of the Zagros will 
find us placing the putative ancestors of the Medes and Persians in their 



historical seats at the same time as the I ndo- Aryans of Mitanni, a situation that 
would probably have been reflected in the historical record. For this reason, it 
has been suggested that the grey ware horizon should be seen as I ndo- Aryan 
rather than Iranian and that for the ancestors of the Persians and Medes we 
must look to other evidence. This may be so, but I think we would do better to 
tie our star to the question of horses and chariots, the only Mitanni evidence 
that is demonstrably Indo- Aryan, rather than to ceramics. It must be admitted 
that at present we lack the necessary intermediary stages of the diffusion of 
horses or chariots from the steppe into northern Syria. We have only points on a 
map - the Trans-Ural steppe, the Pontic, the southern Caucasus, the seal from 
Hissar 1 1 IB. These are not yet arranged into an interpretive framework from 
which we can trace movements, but future excavations are likely to help both fill 
in gaps of evidence and chronology. This, then, is a potentially resolvable 
problem which may provide us with sounder evidence on which to erect 
theories of Mitanni origins. 

The grey wares of northern Iran that have been employed to substantiate a 
Western Iranian presence from the fourteenth century bc onwards have been 
linked genetically to the Gorgan. It will be recalled that about 1800 to 1600 bc 
the proto-urban settlements from Tepe Hissar IIIC eastwards to Namazga and 
Altin Depe witness an almost total collapse and restructuring. The causes of 
this collapse are unknown, although some Soviet scholars argue that steppe 
nomads were an important factor. In the intervening centuries before the 
resurgence of urbanisn in the Iron Age (about 1000 bc), it has been postulated 
that bearers of the grey ware ceramic tradition moved southwest into northern 
Iran. For those who assume a correlation between grey wares and Iranians, this 
spread may be interpreted as the movement of the Western Iranians into their 
earliest historical seats around 1400 bc. Alternatively, this may reflect the 
movement of non-Iranian refugees, especially as the distribution of grey wares 
exceeds the area of the historically attested Iranians and includes non-Iranian 
territories such as Hurrian or Urartian. We must never forget that the grey 
wares can only indicate a trajectory and are not inherently associated with 
Iranian ethnic identity. To reject their utility as markers for tracing Iranian 
movements because grey wares emerged in the Gorgan region as early as 3000 
bc - far too early for an individual Iranian identity - is only compelling if one 
makes the unnecessary assumption that the Iranians had to invent grey wares in 
order to use them. In the southeast Caspian, the culture that emerges in the Iron 
Age, the Dahistan culture, is solidly regarded as Iranian (the land designated 
Varkana (Hyrcania) in early Iranian texts), and it is seen to have emerged out of 
the local Gorgan tradition possibly coupled with steppe influences. It takes no 
great imagination to suggest that these steppe influences may also have drifted 
southwestward with the expansion of grey wares in the preceding centuries. 
The plausibility of future explanations will depend to a great extent on much 
better chronological control of the evidence and a greater consideration of non- 
ceramic remains. What is clear is that our problem is not in discovering possible 
archaeological traces for Western Iranian migrations, but selecting the 



appropriate one. Naturally, by migrations we need not consider a single unique 
event but movements that may have continued for centuries. 

The Eastern Iranians are comparatively secure in their association with 
many of the steppe Bronze Age cultures, although there is great room for 
dispute in selecting specific archaeological evidence to explain the ultimate 
appearance of any individual Iranian group. 

Finally, the arrival of the Indo-Aryans in India itself remains a major 
problem. It is not alleviated by the fact that archaeological research has 
concentrated either in the areas adjacent to the Indus or in Central Asia; only 
now is work beginning in the intervening borderlands, and hence there is a 
plethora of solutions that involve tele-connections between ceramics or metal 
types separated by uncomfortably large distances. We can certainly talk about 
the possibilities of identifying a number of intrusive cultures that may help 
explain Indo- Aryan movements into northwest India although we are not yet 
prepared to erect the type of all-embracing archaeological model that explains 
their relationship with either the western Indo-Aryans or their Iranian cousins. 

The Tocharians 

Tocharian, the easternmost group of the Indo-European languages, is attested 
in Chinese Turkestan from manuscripts dating from the sixth to the eight 
centuries ad. These were recovered initially by the great archaeological 
explorers of the late nineteenth and early twentieth centuries who began to 
acquaint the West with the ancient ruins of this remote area of China. The 
manuscripts themselves were primarily the work of Buddhist monasteries and 
largely comprise translations from Sanskrit, tracts on magic and medicine, and 
occasionally the business transactions connected with the major caravan route 
that passed through the northern region of the Tarim Basin. The language of 
the manuscripts was identified as Indo-European (written in an Indie script 
they were quite easy to decipher) early in this century. The language was named 
Tocharian after the historical Tokharoi who were known to the Greeks to have 
emigrated from Turkestan to Bactria in the second century bc. The full 
arguments for the validity of this designation and other ethnic labels that have 
been pinned on the creators of the manuscripts has been hotly debated for 
decades and comprise a remarkably large percentage of Tocharian scholarship. 
This will not concern us here other than to conclude that there is not a shred of 
linguistic evidence to indicate that the people of historical Tokharistan spoke 
the same language found in the manuscripts of the Tarim Basin well over 1,000 
kilometres to their east. Today few if any would accept that the proper 
designation for these people is Tocharian but as no alternative has ever achieved 
sufficiently wide approval, the earlier name, misnomer if you will, is still 
applied and will be used throughout this work. 

Tocharian is divided into two major dialects. The easternmost is termed 
Tocharian A (or Turfanian or Karashahrian after the two major towns near 
which manuscripts of this dialect were recovered, or Agnean). Tocharian B, the 




Ma ni cisa nos somo nem wnolme tare taka, 
ma ra postam cisa lare masketar-n. 

Cisse laraumne cisse artanye pelke kalttarr solampa sse. 
ma te stalle sol warnai. 

Taiysu palskanoym: sanai saryompa sayau karttses saulu 
warnai snai tserekwa snai nane. 
Yamornikte se cau ni palskane sarsa. 
tusa ysaly ersate, cisy aras ni salkate. 

Waya ci lauke, tsyara nis wetke, lykautka-n pake po laklentas. 
cise tsarwo sampate-n. 

Earlier there was no person clearer to me than you, 
and later too there was none dearer. 

The love for you, the delight in you is breath together with life. 

This should not change for life. 

Thus I thought: with the one beloved will I live well 

lifelong without deceit without pretense. 

The god Karman alone knew this my thought. 

Therefore, he caused dissension and tore from me the heart that belonged to you 
He led you away, separated me and had me partake of all sorrows. 
The joy I had in you he took away from me. 

Tocharian B 


Old Irish 













f rater 














eoh (Old English horse ) 








36 A Tocharian love poem and a short comparison of some haste words in the Tocharian 
languages with those of other Indo-European groups. 

western dialect, is also known as Kuchean after the prominent state of Kucha 
(Kuqa). These dialects are markedly different from one another in both 
vocabulary and grammar, so much so that they have been regarded as two 
separate languages that diverged from one another over a period of 500-1,000 
years despite the fact that when we encounter them they are separated by only 
several hundred kilometres. In addition, there is evidence that Tocharian A 
may well have been on its way to becoming a dead language confined to 
liturgical works by the time of our earliest manuscripts. The evidence from 
Tocharian B, whose texts are found in both western and eastern areas, suggests 
a vigorous language which was at least the vernacular of the religious and civil 
administrative classes of the Kuchean state. If we wish to suggest a date for the 
existence of a common or Proto-Tocharian language, we might expect that it 
had been current in the first millennium bc. The problem becomes far more 
interesting when we turn to the question of where we might expect to have found 
Proto-Tocharian at this time. 



One of the most striking and disarming aspects of the Tocharian languages is 
that their linguistic relationship with both their Indie and Iranian neighbours 
seems to date from a very late period and can generally be attributed to the 
influence of Buddhist missionaries as they pressed eastwards. Thus we find the 
Tocharians with Iranian-speaking Saka and other Iranians to their immediate 
south and w est, possibly also to their north, all showing nothing more than a late 
impact on them. For closer linguistic connections we must look to Europe 
where uniquely similar items of vocabulary and grammar are shared with 
Baltic, Slavic, Greek, Armenian, Germanic and possibly Phrygian and other 
languages. Several features w hich may be of the utmost importance are shared 
with Hittite, Italic and Celtic, which so mesmerized one early scholar that he 
declared Tocharian to be a Celtic language. In Chapter i we encountered the 
famous division between the centum languages of Europe and satem languages 
of Eastern Europe (Baltic, Slavic) and Asia (Indo-Iranian). We can then 
imagine the astonishment that linguists experienced when the Tocharian words 
for hundred (A: ka'nt, B: kante) show ed it to be unequivocally a centum language. 
The implications of all this for determining the initial starting point for the 
Tocharians has been extraordinary. Here they were saddled with similar 
adjectival suffixes with Slavic, a medio-passive ending in -r which had been 
retained in Latin, Irish, Hittite and Phrygian and cognate words shared only 
with Greek. This meant that they must have proceeded from somewhere in 
Eastern Europe and were forced to trek over 4,000 kilometres to take up their 
historical seats. There were even some, obsessed by the fact that the Tocharian 
word for fish was the same as the Germanic, Baltic and Slavic words for salmon, 
who compelled the Tocharians to dine with their European brothers on the 
shores of the salmon-rich rivers of the Baltic before they set off to Turkestan. 

37 The Tocharians of the Tarim 
Basin and sonic of their historic 
(largely Indo-Iranian) neighbours. 



Forced to accommodate the linguistic evidence, archaeologists sought extra- 
ordinarily distant connections between Eastern Europe and the frontiers of 
China. The most serious attempt of this kind was by Robert Heine-Geldern 
who enumerated series of similarities between the metalwork of Europe and 
China about 800 bc upon which he predicated a 'Pontic Migration' from 
Europe across Asia. Before we should assume that the Tocharians are required 
to go to such heroic efforts to explain their presence, we might first seek more 
economical solutions. 

Our knowledge of the Tocharians themselves is almost entirely founded on 
Chinese documents which trace the initial encounters between Han China and 
the barbarians of the western lands from 200 b c until the Tarim Basin became a 
Turkic-speaking region from about the eighth century ad. It should be 
emphasized that the area occupied by the Tocharians was not a desolate refuge 
but rather a series of profoundly important oases along the silk road that led 
from China to the West. The northern Tarim was a major centre for exchange 
and transit, rich in mineral resources, possessing a sound agricultural basis that 
also included horse-raising. The region attracted the constant interest of the 
Hsiung-nu nomads to the north and the Chinese to the east. The history of the 
region is a catalogue of diplomacy and wars between these competing forces 
with the Kucheans who were more successful than their eastern cousins at 
maintaining their autonomy. Tocharians were instrumental in spreading 
Buddhism to China, and the Chinese occasionally tell us something of their 
western neighbours. 

The census for the Early Han (208 bc-8 ad) indicates that there were about 
100,000 people in the Kuchean state, over one-fifth of which were military. 
Later documents describe the Kucheans as a sedentary population practising 
mixed agriculture, frequently dining on peacocks and, like good barbarians 
everywhere, given to excessive drinking. They practised skull deformation with 
a board to flatten an infant's head and wore their hair cut at the neck except for 
the king who wore his long and tied up with a band - * practice illustrated on a 
wall painting from the region. The Kucheans armed themselves with bows, 
swords, long spear and armour, and they cremated their dead. Their cultural 
practices were regarded as similar to their eastern cousins and we know of 
marriages between the royal families of the Tocharian A and B regions. On the 
other hand, the Chinese contrasted the culture of the Kucheans with both the 
Hsiung-nu and the Wu-Sun, their nomadic neighbours. 

When we confine our attention to Chinese historical documents, our image 
of the people who produced the Tocharian manuscripts is geographically 
precise but it does have severe limitations. We have every reason to believe that 
Tocharian was not the only language spoken in these states, and we can only 
assume that its speakers comprised at least the monks and civil authorities, 
excepting, naturally, the Chinese. Secondly, the Chinese sources provide us 
with the political states of Kucha, Karashahr and Turfan but no larger ethnic 
entities. These have nevertheless been frequently proposed. The Wu-Sun, for 
example, occupied the territory immediately to the north of the Tocharians and 



are described in Chinese sources as having red hair, blue eyes and resembling 
monkeys. This pejorative, proudly seized upon by western scholars as 
indicating a Europoid population, does indicate a strong possibility that the 
Wu-Sun spoke either an Iranian or Tocharian language. The Tocharians 
themselves are depicted on wall paintings as having red hair and green eyes. 

Another tribal confederacy, the Yueh-chih, plays an especially prominent 
role in the Tocharian problem. They first emerge in northwest China in the 
second century bc when they were defeated by the Hsiung-nu confederation of 
tribes (probably mixed racially and linguistically) who rendered the skull of the 
Yueh-chih king into a drinking cup. They fled westwards into the territory of 
the Wu-Sun only to be driven onwards again further to the west, ultimately to 
be identified with the tribes that settled historical Tokharistan. The territory of 
the Yueh-chih is seen by many also to have included the Tarim Basin and hence 
they have been credited as Tocharian speakers. This is a bit too much of the tail 
wagging the dog since we may well expect that the major tribal groups were very 
mixed and while Tocharian may have been once spoken by some members of 
the Yueh-chih, there is no reason to assume that the Yueh-chih were a 
Tocharian linguistic entity. In any event, in the first centuries AD the Chinese 
regularly differentiate the Tocharian-speaking states of the Tarim (Kucha, 
Karashahr, Turfan) from the Yueh-chih, Wu-Sun and Hsiung-nu. This was 
probably a distinction made on economic (nomads versus agriculturalists) and 
administrative principles and certainly not on linguistic ones. Hence the Wu- 
Sun or Yueh-chih - or at least some member of their tribes - cannot be 
demonstrated not to have spoken Tocharian, but despite occasional attempts to 
squeeze Tocharian etymologies out of their tribal names, there is really no 
convincing evidence one way or the other for what language they actually did 
speak during the first millennium bc, although Iranian and/or Tocharian 
would be fair assumptions. 

Having indicated something of the scope of our ignorance, we can now see 
how far we can push towards a solution to the Tocharian problem. Our first 
unequivocal evidence for Tocharian speakers is no older than our earliest 
documents, about 600 ad. If we wish to assume that their presence in the Tarim 
Basin extends earlier, we must rely on indirect evidence. The close association 
of the documents with the Buddhist mission may permit us to assume that the 
Tocharians were present at least since the appearance of Buddhism in the 
region, variously dated from the second century bc to the first century ad. 
Moreover, Chinese accounts of the history of the Tarim Basin do not suggest 
that a new people had altered the ethnic identity of the Tarim states from the 
second century b c onwards. We must be cautious here, however, since there are 
ample instances of Hsiung-nu conquests, political usurpation, and intermar- 
riage between the royal family of Kucha and its neighbours, all of which could 
disguise an ethnic intrusion. But there is certainly no compelling reason to 
assume that the Tocharians had not been present since at least the second 
century bc. Beyond this, historical testimony is totally silent, and the 
archaeological evidence for the Tarim Basin becomes exceedingly dim until the 



Neolithic (4000-2000 bc), when we find evidence for the painted wares of the 
Yang-shao and the monochrome wares of the Longshan horizon. In both cases 
we are talking of cultures that later produced the distinctly Chinese societies of 
the Shang and Zhou periods. If the Tarim Basin was essentially a western 
extension of these Chinese cultures during the Neolithic, then it may be 
assumed that the Tocharians must have entered the region after this period. We 
may thus seek the earliest Tocharians in the very broad period of between 2500 
and 200 BC. 

To proceed any further we must return to the problem of the linguistic 
relationships of Tocharian. We have already seen how its close links with the 
European languages generated a model of a European homeland with a 
subsequent migration of over 4,000 kilometres. This migration seems to be set 
in general to the first millennium bc. On purely logistic grounds such a 
migration cannot be excluded since we have the historical examples of Huns, 
Alans and other steppe nomads. Yet it must also be admitted that any 
movement from the west to the east along the steppe would appear to be 
running counter to the general east-west current of steppe populations of the 
first millennium bc and later. In addition, the movements of the historical 
nomads involved an accretion of intermediate populations so that by the time 
the Huns, for example, burst into Central Europe they were an amalgam of 
Turkic, Iranian and Germanic-speaking peoples. Tocharians moving through 
thousands of kilometres of what we may expect to have been Indo-Iranian or 
more specifically Iranian territory should have emerged with a far greater 
accretion of Iranian influence than we find. In short, there are serious reasons to 
doubt that the Tocharians achieved their position on the eastern flank of the 
Iranians as late as the Iron Age. 

An alternative model of Tocharian's relation to the European languages 
offers some hope of a solution. It weighs the similarities shared between 
Tocharian, Celtic, Italic and Hittite as essentially archaic features inherited 
from the Proto-Indo-European language at a very early period. These 
grammatical features were then replaced in later Proto-Indo-European by new 
forms that spread among the ancestors of Greek, Armenian and Indo-Iranian, 
but not to what had then become the outer periphery of the Proto-Indo- 
European continuum - the ancestors of Celtic and Italic on the west, Hittite and 
possibly Phrygian to the south, and Tocharian on the east. 12 In the eastward 
expansion of Indo-European languages, Tocharian preceded Iranian into 
Turkestan and was later engulfed by Iranian-speaking Saka to the south, 
Sogdian and others to the west, and, if Iranian river names in the Minusinsk 
Basin are included, also to their northeast. The occasional lexical similarities 
with some of the other European languages may then be dismissed as the chance 
occurrences that the historical linguist must expect of his material. 

The archaeological evidence of the Tarim Basin is still far too poorly known 
to permit us to test our linguistic model archaeologically. Still, if we paint our 
narrative with an exceedingly broad brush, we may see something of a hint of an 
explanation. Earlier we saw how the Andronovo culture of the Asian steppe 



3# The Tocharians may have 
found their immediate origins in 
the eastern Andronovo or 
Afanasievo cultures. 

seemed to make an excellent correlation with the later appearance of Eastern 
Iranian peoples. But we need not be so perverse as to demand an exact 
correlation between an archaeological culture, especially one as vague as the 
Andronovans, and a single linguistic group, and it is entirely possible that the 
ancestors of the Tocharians lurked behind some of those Andronovo variants 
that appear in the southeastern area of its distribution. This would include 
Tadzhikistan and Kirghiziya to the w est of the Tarim Basin where Andronovo- 
related sites begin to appear by at least 1400 bc. Although separated from the 
Tarim Basin by mountain ranges, these people were hardly strangers to high 
altitudes, especially those w ho occupied the Tien Shan region. One cemetery at 
Arpa, for example, is found at an altitude of 2,800 metres above sea level and is 
situated less than 500 kilometres west of demonstrably Tocharian-speaking 
territory. It may only be coincidental that this region does not practise the usual 
inhumation of the Andronovo cultures, but rather, like the Tocharians, 

For those who require some form of symmetry between language and 
archaeological culture, there is yet another possibility. Prior to the Later Bronze 
Age appearance of the Andronovo culture across the Central Asian-west 
Siberian steppe, there appeared an Eneolithic culture whose boundaries were 
apparently confined to the Minusinsk-Altai region, 1,000 kilometres to the 
north of our Tocharians. This Afanasievo culture, dating to the third 
millennium bc, possesses many of the attributes that we often demand of any 
Indo-European culture - domestic horse, basic metallurgy, and possibly 
wheeled vehicles. What makes this culture so important is that it stands without 
any clear connection with the cultures to its north or east yet possesses clear 
analogues in ceramics, lithics, burial practices and physical type with the west, 
specifically with the Volga-Ural region. We will have very good cause to 
examine this culture in Chapter Eight; for now it suffices to observe 
that we have evidence of a possible eastward expansion from a territory 



producing other Indo-European speaking peoples at a date prior to the 
expected spread of the Indo-Iranians. It is possible that this represents the 
archaeological correlate for the eastward spread of the peripheral Indo- 
European languages among which Tocharian has fortuitously survived. 


By virtue of its early development of writing, Asia offers us information about 
the early expansion of the Indo-Europeans that is much more difficult to 
acquire or substantiate elsewhere. We have seen that whenever we find 
decipherable Bronze Age documents, they attest to the intrusive character of 
the Indo-European presence in Asia. Hittite replaces Hattic (and Assyrian), 
Armenian absorbs the peoples and languages of the Indo-European-speaking 
Luwians and the non-Indo-European Urartians, Iranian expands over the 
earlier territory of Elam, and if we subscribe to the most acceptable approaches 
to the Indus script, Indo-Aryan made an impressive expansion over much of 
the vast non-Indo-European populations of India. It can hardly be doubted 
that from when we first acquire written documents in the Bronze Age, we find 
abundant evidence for migrations and linguistic replacement. In central 
Turkey alone, for example, the non-Indo-European Hattic is replaced by a 
series of Indo-European languages - first Hittite, then Luwian, then Phrygian, 
then a Celtic language that gives its name to the ancient province of Galatia, 
and finally Greek - only to be absorbed once again by the wholly different non- 
Indo-European language of the Turks who immigrated from the east. The 
history of the Near East also documents the expansion of Semites, Hurrians and 
others through the course of the Bronze and Iron Ages. All of these examples 
provide a salutary reminder to archaeologists that populations often shifted 
their boundaries no matter how poor our ability to trace such movements in the 
archaeological record or how much the model of migration conflicts with 

jg Early Neolithic sites of 
Southwest Asia and the 
distribution of non-Indo- 
European languages during the 
third millennium BC. 



current approaches to culture change that stress local processes rather than 
folk-movement. It should be emphasized that such movements are in no way 
confined to state level societies (as anyone familiar, for example, with the 
distribution of the major aboriginal Indian languages in North America well 
knows - such as the Athapascans of northwest Canada and their southern 
linguistic relations, the Navaho and Apache, in the American Southwest). Folk 
movements may not have been exactly daily occurrences but they are amply in 
evidence and one need not stubbornly adhere to what Christopher Hawkes has 
recently termed the ideology of 'immobilism' to be interested in the processes of 
cultural change. 

We may briefly reflect on the obvious demise of a number of the languages 
that here preceded the intrusion of the Indo-Europeans into Asia. The Indo- 
Europeans account largely for the total disappearance of at least three separate 
linguistic entities, perhaps language families - Hattic, Hurrian-Urartian and 
Elamite; the Semites must probably explain the ultimate disappearance of 
Sumerian. When we estimate the territory of each of these language families 
they approximate an area roughly equivalent to that of Germany or Poland or 
Japan. Naturally, some seem to occupy a much smaller area, such as the 
Sumerians, while others, such as the Semites, probably covered vast expanses 
during the Bronze Age. Whether or not these are fair estimates for the 'typical' 
territories of Bronze Age language families about 3000-2000 BC is difficult to 
say, especially as much of our sample derives from primarily urban populations 
which we will not encounter elsewhere. Nevertheless, the area traditionally 
defined as the Turkic homeland before their expansion in the early Middle 
Ages is of a similar order of magnitude. We may wish to keep the scale of these 
linguistic territories in mind when we begin to examine the Indo-European 
homeland itself. 

As the evidence reviewed so far indicates the intrusive nature of the Indo- 
Europeans in most of Western Asia, this provides us with reasonable grounds 
for excluding these regions from the territory of the earliest Indo-Europeans. 
With the emergence of our earliest historical documents in the third 
millennium BC we find non-Indo-European populations occupying central and 
eastern Anatolia to the Caucasus and Caspian, the southern region of historical 
Palestine, and much of the Zagros region including all of Mesopotamia. These 
areas comprise the three main centres of incipient agriculture during the 
'Neolithic Revolution' (9000-6000 bc) in Southwest Asia. Consequently, it 
would appear to be highly unlikely that we should associate the Indo- 
Europeans with these earliest farming communities in Southwest Asia and their 
immediate expansions. One might, of course, suggest some complicated reflux 
model where the earliest Indo-Europeans, for example, the Anatolians and 
Armenians, were driven from their homes by early Hattic or Hurrian 
expansions and then reclaimed these territories in the third to first millennia bc. 
This, I would think, not only requires special pleading but is inherently 
unlikely. We have too much evidence for unanalyzable place and personal 
names across Anatolia to suspect anything other than a non-Indo-European 



substrate. Moreover, if the Indo-Europeans had emerged in this very region, as 
Colin Renfrew has recently suggested, one might have expected far greater 
similarity between Proto-Indo-European and the languages of its non-Indo- 
European neighbours. For these and other reasons which we will examine later, 
the great majority of linguists would seek the ancestors of the Indo-Europeans 
of Asia somewhere to the north of their later historical seats. But before we can 
take up this trail again, we must first look westward and examine the evidence 
for the earliest Indo-Europeans of Europe. 

6 5 



The Indo-Europeans in Europe 

In our school days most of us were brought up 
to regard Asia as the mother of the European 
peoples. We were told that an ideal race of men 
swarmed forth from the Himalayan highlands, 
disseminating culture right and left as they 
spread through the barbarous West. 

Joseph Ripley, 1900 

The evidence for the earliest Indo-Europeans in Europe differs in several 
fundamental ways from most of that which we have encountered in Asia. With 
the exception of Greek, most of our earliest testimony of Indo-European 
speakers is considerably later than that of their Asian cousins. Writing appears 
later in Europe and, consequently, we cannot hope for the direct testimony of 
the different Indo-European groups until comparatively recent times. This 
alters the nature of our evidence since we invariably find that the different 
Indo-European groups predate the historical record in their respective 
locations, making it more difficult to determine if particular Indo-European 
languages are intrusive, and, if so, when they arrived. Consequently, we are 
often thrown back on other lines of evidence, particularly archaeological and 
linguistic coupled with occasional forays into para-history, the relations of 
Greek and Latin authors to their poorly known neighbours. In reviewing the 
earliest Indo-Europeans we will begin in the Aegean and then move through 
the Balkans and Eastern Europe, completing our survey with the Indo- 
Europeans of Western Europe. 

The Greeks 

As with Anatolian and Indie, the Greek language is first attested in the Late 
Bronze Age. Our earliest evidence is to be found among the over 3,000 clay 
tablets recovered from Knossos on Crete, and from Mycenae and Pylos on 
mainland Greece. The majority of these tablets, generally economic records for 
the palace economies of the Mycenaean and Late Minoan civilizations, are 
w ritten in a syllabic script known as Linear B. The Linear B tablets date from 
about the thirteenth century B c and most linguists accept that they were written 
in a primitive form of Greek, frequently termed Mycenaean. With the collapse 
of the Mycenaean palatial civilization by the twelfth century bc, written 
evidence for Greek disappears until the introduction of the alphabet sometime 
between 825 and 750 bc. From this time onwards, there are both inscriptions 


and the compilation of the|preat Greek oral epics such as the Iliad. In the 
subsequent 2,500 years the Greek language has undergone marked changes to 
evolve into Modern Greek which is currently spoken by approximately eleven 
million people both in Greece and in a few of its remnant colonies. 

The investigation of Greek origins has long been a subject of intense interest 
and there is a sufficient number of wholly contradictory theories concerning the 
'coming of the Greeks' to banish any fantasies of universal agreement among 
either linguists or archaeologists. Nevertheless, there is a course we can steer 
that will at least follow the tack taken by the majority of scholars and we can note 
where others depart from this. 

Why must the Greeks 'come' from elsewhere? In attempting an answer to 
this question we must admit that we lack the historical documentation that 
served us in Asia and, consequently, we cannot positively demonstrate a pre- 
Greek population on the evidence of contemporary testimony. 13 Yet it has long 
been argued that a pre-Greek population declares its existence loudly enough in 
the Greek language itself. There are two types of evidence. 

Many would agree that a substantial portion of the Greek vocabulary 
pertaining to the specific environmental character of the Mediterranean cannot 
be explained as the Greek outcome of inherited Indo-European words. These 
include such plants as the vine, fig, olive, hyacinth, cypress, laurel, marjoram, 
chickpea, chestnut, cherry and parsnip. Among the animals are the ass, the wild 
ox and the beetle. Items of material culture are similarly non-Greek: metal, tin, 
bronze, lead, jar, pail, oil flask, sword, javelin, cornice, coping, chamber, bath 
tub and brick. Certain political or social concepts, basic to Greek society, are 
expressed with what are usually taken to be pre-Greek words. These include 
the word for king (basileus, Mycenaean qa-si-re-u)^ as well as slave (doulos, 
Mycenaean do-e-ro) and concubine. Neither do the most prominent heroes of 
Greek epic - Odysseus, Achilles, Theseus - nor many of the Greek divinities - 
Athene, Hera, Aphrodite, Hermes - bear obviously Indo-European Greek 
names. 14 

A second line of evidence derives from place names found in Greece. Many 
of the most important names cannot be explained by the Indo-European core of 
the Greek language. Either the roots of the names are meaningless in terms of 
Greek, or the names are constructed employing suffixes which signal that they 

Transliteration pu - ro i-je-re-ja do-e-ra e-ne-ka ku-ru-so-jo i-je-ro-jo woman14+ 

Greek jiuAxx; i£peux<; 5oeXm eveko xpwoto lepoio woman 14+ 

Transliteration Pyios hiereias doelai heneka khrysoio hieroio woman 14+ 

Translation Pylos: priestess' slaves on account of gold sacred 14+ women 

40 An inscription in Linear B. 



are not originally of Greek construction. These include major toponyms such as 
Corinth, Knossos, Salamis, Larisa, Samos and even Olympus and Mycenae. In 
an early interpretation of this evidence, Joseph Haley and Carl Blegen observed 
that many of the names and suffixes were also to be found in Anatolia and that it 
was probable that they expanded from that region into Greece during the Early 
Bronze Age (3000-2000 bc) when there was a general cultural uniformity over 
much of the Aegean. Whether all the details of this hypothesis can withstand the 
more critical scrutiny in current archaeological and linguistic circles is 
doubtful. Yet the linguistic evidence taken as a whole does indicate that the 
Greeks did borrow a considerable number of elements from a non-Greek 
language. 15 The vocabulary suggests that these borrowings were not wholly 
random, but rather tend to focus on words that a population intrusive into a new 
land might be expected to adopt from the previous inhabitants. At least some of 
the loan words borrowed into Greek would seem to have been derived from a 
culture familiar with a level of metallurgy (copper, bronze, tin) that existed no 
earlier than the end of the fourth millennium bc. The logical consequence of all 
this is that the Greeks are not native to Greece but were the product of Indo- 
European intruders who superimposed themselves on an earlier Bronze Age 
population. This, some would say, is further confirmed by the Greeks' own 
historical tradition which indicates that they had absorbed a number of earlier 
non-Greek populations, some of whom bear obvious non-Indo-European 

The concept of a pre-Greek population is commonly agreed upon; the 
interpretation of the pre-Greek elements in the Greek language, however, is 
not. In identifying the nature of this pre-Greek element there are two major 
schools of thought with opinion scattered from one extreme to the other. One 
group terms the pre-Greek element 'Mediterranean' or 'Aegean' and sees it as 
the remnants of the non-Indo-European language(s) that were once spoken in 
Greece and perhaps more widely in the Mediterranean. The second school of 
thought admits that, while some of the words are indeed non-Indo-European, 
many others, nevertheless, are Indo-European. Some identify the Indo- 
European element as Luwian, especially on the basis of place names which 
occur both in Anatolia and Greece. Others opt for an otherwise unattested 
Indo-European language, closely related to some of the poorly attested Balkan 
languages, which is generally known by the name of Pelasgian, one of the 

In earlier times there were two races living in Greece: the Pelasgians, 
who never left their original home, and the Hellenes (Greeks), who 
frequently migrated. What language the Pelasgians spoke I cannot 
say exactly. If it is proper to judge from those of the Pelasgians who 
still survive, they spoke a non-Greek language. If that was true of the 
whole Pelasgic race, the Attic nation must have learnt Greek at the 
same time that they became Hellenized. 

41 Herodotus s account of the Pelasgians reflects Greek historical tradition that the Greeks 
absorbed an earlier non-Greek population. 



'autochthonous' peoples of Greek historical tradition. General linguistic 
opinion has neither been particularly impressed nor even kind to the 
Pelasganists, and most would still regard the pre-Greek vocabulary to be largely 
unanalysable in terms of Indo-European with perhaps a grudging admission 
that some non-Greek words can pass muster as Indo-European, possibly 
Luwian or an anonymous Pre-Hellenic language, while others may be later 
Semitic loan words. Consequently, the linguistic evidence does seem to indicate 
that the Indo-Europeans who eventually emerged as the Greeks were 
essentially intrusive into Greece and that they mixed there with a non-Indo- 
European population and possibly with earlier Indo-European-speaking 
peoples as well. 

Acceptance of the decipherment of Linear B as Greek leads us to the 
inescapable conclusion that Greeks must have been present in Greece by about 
1300 bc when we find the earliest texts. These were discovered in the palace of 
Knossos on Crete and are generally interpreted as the product of an intrusive 
Mycenaean power which adopted the earlier Cretan writing system, known as 
Linear A, which was already flourishing by 1 700 B c. Neither the language of the 
Linear A script nor the culture of Minoan Crete are regarded as Greek and its 
population is generally believed to have been thoroughly native at least from the 
Early Bronze Age, if not from the Neolithic. There is also sufficient evidence to 
indicate an intrusion of a Mycenaean warrior aristocracy into Crete about 1450 
BC to permit archaeologists to postulate a genuine political take-over of the 
Minoans by the Greek-speaking Mycenaeans. The Mycenaeans appear to have 
adapted a near relative of the Minoans' Linear A script to suit their own 
language. The sheer clumsiness in expressing the Greek language by either the 
Linear A or Linear B syllabaries lends support that they were not originally 
invented by Greek speakers. 16 It is also clear from this that we cannot follow the 
trail of Greek origins on Crete but must look to southern Greece and the origins 
of the Mycenaeans. 

The great Myceanaean citadels began about 1400 bc, and there is no reason 
to dissociate the people who built them from those who produced the Linear B 
texts. The Mycenaeans themselves are normally derived by way of a processual 
jump in social complexity from the earlier Middle Helladic culture whose roots 
extend back into Early Helladic III, about 2200 BC. There is, however, a sector 
of opinion which does not credit the sudden rise of Myceanaean chiefdoms to 
purely internal evolution but seeks an exterior stimulus. The stimulus is 
identified as a small body of warlike intruders who introduced the horse and 
chariot, new weaponry such as swords and the body shield, and status burials 
under a tumulus. These appear during the Middle Helladic period and 
culminate with the tumulus burials at Marathon which include rich burials that 
extend from the Middle Helladic into the Mycenaean period. Moreover, a 
horse burial in one of the Marathon tombs has been advanced as further 
evidence of Indo-European intruders. These tumulus burials are best known 
from western Greece and are held up in contrast to the exceedingly poor 
intramural graves that typify most Middle Helladic sites. A possible external 



42 The distribution of tumuli in Greece 
an J Albania in the Early anil Middle 
Bronze Ages. 

43 7 he distribution of apsidal houses, 
clay anchors and stone battle-axes in 
Early Bronze Age Greece. Some suggest 
that these spread from Southeast Europe 
with the 'coming' of the Greeks. 

source for them is provided by roughly contemporary tumulus burials to the 
north, especially in Albania, for example, Pazhok and Vodhine. Consequently, 
a few archaeologists see evidence for an intrusion from the north during the 
Middle Bronze Age, particularly associated with burials of a higher status than 
the native burials, and also directly linked with the emerging ranked society of 
the Mycenaeans. The evidence hangs to a large extent on what weight one 
attributes to the tumulus burial itself as an ethnic marker for there are really few- 
other items associated with these burials that weld them together into an 
intrusive culture. The sword and body shield, for example, only begin to appear 
in number at the very end of the Middle Helladic period in the shaft graves at 
Mycenae. The only other element, perhaps, is the horse, which also makes an 
appearance on Albanian sites in the Middle Helladic period or earlier. There is 
some doubt regarding the Marathon horse sacrifice since it now seems, in fact, 
to have been an insertion from a much later period. 

A far more widely accepted solution to the coming of the Greeks is the 
discontinuity that occurs between Early Helladic II and Early Helladic III, 
about 2200 bc. The relevant evidence includes the destruction and abandon- 
ment of Early Helladic II sites, changes in architecture especially with 
reference to the appearance of houses with an apsidal end, burials within 
settlements (although our evidence for this does not begin until the Middle 
Helladic period), perforated stone 'battle-axes', clay 'anchors' and the 
incessantly discussed Minyan ware. Following this break, mainland Greece 
appears to revert to a much simpler agricultural society, generally regarded as 



retarded compared with its neighbours on Crete who escaped the Early 
Helladic III destructions. Since all cultural development subsequent to the end 
of Early Helladic II appears to suggest almost monotonous continuity, it has 
been accepted by many that the new changes in culture are most likely to be 
attributed to those intrusive Indo-Europeans who later emerge as the earliest 
Greeks. Generally, the source of this intrusion is held to have been either from 
Anatolia or the Balkans. The former perhaps offers better ceramic parallels 
(which may ultimately be only the product of specialist craftsmen employing 
the fast wheel) while the Balkans, particularly the Ezero and Baden cultures, 
offer some of the earliest examples of apsidal houses. It is argued that these 
subsequently appear in western Anatolia and then slightly later in southern 
Greece. Stone battle axes and possibly a number of other traits are also tied to 
the Balkans. 

Finally, one should mention the transition from the Late Neolithic to the 
Early Bronze Age itself since we have already seen that it is precisely in this 
period that some see movements from the Balkans into northwest Anatolia. The 
emergence of Early Helladic I about 3000 b c is still very poorly understood, and 
the degree of continuity from the Late Neolithic versus intrusion is by no means 
settled. This represents the earliest potential Indo-European invasion in the 
minds of most scholars, although at least one Pelasganist has opted for a Late 
Neolithic invasion to explain their presence in Greece. 

Reconciliation of all these different theories seems out of the question, 
although acceptance of each hypothesis as evidence of an actual migration gives 
more than enough scope to slip Greeks, Luwians and anyone else in as well. The 
current state of our knowledge of the Greek dialects can accommodate Indo- 
Europeans entering Greece at any time between 2200 and 1600 bc to emerge 
later as Greek speakers. It is the archaeological evidence which really sits in the 
balance and must be more closely examined on a regional basis. The main 
evidence for Early Helladic II/III destruction, for example, is in southern 
Greece and the putative invaders are seen to derive from central Greece and not 
directly from outside Greece. Consequently, links between new cultural 
features in Greece and either the Balkans or northwestern Anatolia (other than 
some ceramic links and possibly apsidal houses) are yet to be built up in the 
necessary stepping-stone fashion that would ensure acceptance. Some archae- 
ologists would argue that there are sufficient numbers of broad trends to see a 
pattern of intrusion and migration; others would dismiss the evidence and, even 
when granting some claims to immigrating populations, would see them as too 
few to have effected the linguistic change postulated by our model of the 
linguistic history of Greece. As we will see many times again, and have cause to 
examine more closely in a later chapter, there are no hard and fast rules for 
defining how much evidence is required to demonstrate an intrusive population 
resulting in the spread of a new language. For our present purpose, it is enough 
to note that there are plausible models of intrusions that are temporally 
congruent with our linguistic models; whether they are based soundly on the 
archaeological evidence is a different question. 



The Thracians 

The term 'balkanization', the breaking up of a territory into small hostile states, 
although of recent creation is also an apt description of the Indo-European 
groups which first emerged into history in the Balkans. These comprised the 
larger 'ethnic' units of the Thracians, Dacians, Getae and Illyrians, which were 
in turn subdivided into countless smaller tribes and tribal configurations. It is 
the precise geographical positioning of these tribes, their ascription to one of the 
larger ethnic groups and their possible association with peoples bearing the 
same names in neighbouring Anatolia that frequently occupy the attention of 
the modern scholar. 

It is a sad irony that the people described by Herodotus as the greatest and 
most populous on earth (after the Indians), the Thracians who occupied the 
eastern half of the Balkans, have left no modern descendant of their language. 
Rather, we must content ourselves with but two inscriptions of disputed 
interpretation; glosses, especially on the names of plants since the ancient 
Greeks regarded the Thracians and Dacians as masters in healing herbs; and an 
abundance of onomastic data - the names of peoples and places. From the more 
reliable etymologies, a number that would not exceed fifty, and their 
grammatical endings, we can safely maintain that the Thracians spoke an Indo- 
European language and say something of its phonetic structure. The Dacian 
language, spoken by contemporary populations north of the Danube in 
present-day Romania, offers slightly less evidence, with some twenty-five 
words that can be given respectable Indo-European etymologies. With such 
meagre evidence it has been impossible to determine whether Dacian and 
Thracian represent two distinct languages or markedly different dialects of the 
same language. Certainly it is odd that the standard suffix in Dacian indicating a 
town, -dava, is not reflected in any of the three Thracian words for town, village 
or fort {-bria, -para, and -diza). 

Both the Thracians and the Dacians achieved the status of impressive and 
powerful kingdoms before falling to the Romans. The Thracian Golden Age 
may be set to the flowering of the Odrysian state which began in the fourth 
century bc, while Dacia looks to the reign of Burebista in the first century bc, 
who achieved unity across much of the northeast Balkans. Our historical 
sources for these people, of course, run earlier, with the Thracians appearing as 
enemies of the Greeks by the mid-seventh century bc when they challenged 
their colony at Thasos, and still earlier they achieved their hostile reputation 
with the Greeks by siding with the Trojans in the Iliad. There are few if any 
who would contest the existence of the Thracians by the Iron Age. How much 
further back can we push them? 

Attempts to seek the origins of Thracians, Dacians or Illyrians as we will see 
later, involve the concept of ethnogenesis among leading East European 
archaeologists and linguists. Adopting the term ethnos, a people with a common 
language and shared customs, most East European archaeologists argue that the 
creation of the various Balkan ethnoi involved a long complicated process. This 
included both local Balkan continuity and frequently intense interactions and 



influences from neighbouring peoples to produce the ethnic groups that one 
first encounters in the historical record. They argue that a basic continuity 
stretching from the Iron Age back into the Bronze Age cannot be denied and 
that this period sees the gradual evolution of a Thracian ethnos, including 
language, from earlier Indo-European components. This thread of basic 
continuity is normally extended back to the beginning of the Early Bronze Age, 
sometime prior to 3000 b c. Beyond this one encounters a major discontinuity in 
the archaeological record which many attribute to the earliest influx of Indo- 
European speakers into the Balkans. 

We will examine the evidence for this invasion in Chapter Eight as any 
movement of putative Indo-European speakers in the fourth millennium bc 
concerns not only the origins of an individual Indo-European-speaking group 
but that of the Proto-Indo-European community itself. One important 
example of this discontinuity in the eastern Balkans will suffice for now. 

Archaeologists can mark the gradual development of the earliest Neolithic 
settlement in the east Balkans by way of the large tell sites that were produced 
by generations of continuous settlement from the sixth millennium onwards. 
Such sites include Karanovo, the cultural yardstick of the east Balkans, which 
shows six major phases of evolution from the earliest Neolithic through the 
Eneolithic cultures of the late fifth millennium. Neighbouring sites also reflect 
similar sequences of local evolution. But following Karanovo VI there is 
widespread abandonment of these tell sites and only a few reveal evidence of a 
later Karanov o VII phase of resettlement. This last phase has little to do with 
any of the previous cultures of the tells and is regarded by many to have been the 
product of intrusive populations infiltrating the lower Danube region from the 
Pontic steppe. These intruders initiate the Early Bronze Age of the Balkans 
and, in the opinion of many, they also introduce a very early form of Indo- 
European language among the native populations of southeast Europe. After a 
fusion of native and Indo-European intruder in the Bronze Age, the major 
ethnic groups of the Balkans eventually began to crystallize in the Late Bronze 
Age to emerge in the early historical period as the Thracians. 

The Illyrians 

Next to Thracian the other great Balkan language was Illyrian which, like its 
eastern neighbours, comprised a series of tribes or tribal confederations whose 
linguistic identity is seldom entirely certain and which offers ample scope for 
historian, linguist and archaeologist to debate the problems of Illyrian 
ethnogenesis. There is, however, one major difference between the Illyrians of 
the west Balkans and the Thracians: it is at least possible that a modern 
descendant of one or more of the Illyrian languages is still spoken. Albanian, a 
language spoken today by approximately five million people, is situated in 
territory earlier attributed to the Illyrians. This permits many linguists to 
surmise that it is probably a descendant of the earlier Illyrian language although 
much affected by Latin, Greek, Slavic and Turkish. We must insist that this is a 



44 An Albanian pJem. These lines contain 
basic words inherited from Proto-Indo- 


Ne grure kish rene draperi, 
ish thare bari ner ara, 
pelciste per uje gjarperi 
dhe binin mullareve zjarre . . . 

European. For example, the Albanian word 
for fire zjarr is from Proto-Indo-European 
*g xv her- which yields Greek thermos, Latin 
formus, Armenian jer, all meaning 'warm', 
Old Irish gorn 'fire', and Russian gorn 

Upon wheat had fallen the sickle 
was dried the hay within fields 
sloughed off through water the serpent 
and spring in haystacks fires . . . 

'hearth'; uje 'water' is from Proto-Indo- 
European # wedor which yields Sanskrit 
udan, Greek hudor, Umbrian utur, Old 

Church Slavonic voda, and English water. 
There are also later loan words such as 

Northwest Greek drapanon 'sickle 1 borrowed 
into Albanian as draperi. 

surmise and not a certainty since there is little if anything to provide clear 
linguistic proof of the connection between Illyrian and Albanian: the modern 
language is known from written records only from the fifteenth century ad 
onwards, by which time it had undergone massive attrition in vocabulary to the 
above four languages. On the other hand, the linguistic evidence for Illyrian is 
pitifully small except for personal and place names. Otherwise the linguist must 
be satisfied with a few paltry glosses such as the Greek identification ofsabaia as 
a type of beer, or that the Illyrian word for fog was rhinos, all of which provides 
scant basis for discussing its relationship with Albanian. Without actual Illyrian 
texts, we must content ourselves with regarding Albanian as an Illyrian 
language as merely a probable assumption. 

Through a long series of wars the territory of the Illyrian tribes was finally 
incorporated into the Roman empire by the first century bc from which time 
onwards it became a truly integral part of the Roman state, a recruiting area for 
the Roman army and the home or birthplace of emperors such as Diocletian and 
Constantine as w ell as scholars such as Saint Jerome. As a major naval power of 
the Adriatic, dismissed as pirates by its Greek and Roman neighbours, the 
Illyrians could not fail to come into open conflict with those around them. 
Greek colonies had been established in the south of Illyrian territory as early as 
the seventh century bc and the names of tribes such as the Paeones and 
Dardani, later recorded in Illyrian territory, are mentioned by Homer as allies 
of the Trojans. They were also traditional opponents of the Macedonians and 
the immediate forefathers of Alexander the Great developed their military skills 
in their frequent conflicts with their Illyrian neighbours. The Illyrians also 
encountered the expansion of the Celts into southeast Europe during the fourth 
century bc. 

Radoslav Katicic's study of personal and place names within the territory 
traditionally assigned to the Illyrians permits us to locate them in two major 
groups: a southeastern group which occupied Dalmatia as far south as Epirus, 
and a central group of names stretching from Dalmatia over western and 
southern Bosnia. This is the safest core region to label as Illyrian-speaking and k 
it is the one with which we will concern ourselves. 



Both Yugoslav and Albanian archaeologists are generally agreed that they 
can trace a fairly direct continuity of culture in this region back to the 
beginnings of the Early Bronze Age. The pivotal site of Albania, Maliq, offers a 
clear succession of cultural development, clearly influenced by foreign 
connections (especially with Bronze Age Greece) back to the Early Bronze Age 
levels of Maliq Ilia. Here, in marked contrast to the earlier Maliq lib levels, 
there seems to be an abrupt break signalled by the appearance of a new culture 
with cruder ceramic types, frequently including double-handled vases and 
single-handled cups, with no connection with earlier Eneolithic levels. 
Furthermore, the appearance of tumulus burials in Bronze Age Albania is also 
regarded as intrusive and derived from the influx of steppe pastoralists of which 
we will speak later. Bosnia is similarly regarded by Yugoslav archaeologists as 




Amantini y Dan ube 

Deuri n M . 
^Delmatae Daesit,ates Dindari 



They say that the country received its name 
from lllyrius, the son of Polyphemus; for the 
Cyclops Polyphemus and his wife, Galatea, 
had three sons, Celtus, lllyrius, and Galas, all of 
whom migrated from Sicily, and ruled over the 
people called after them Celts, lllyrians and 
Galatians. Among the many myths prevailing 
among many peoples this seems to me the 
most plausible, lllyrius had six sons, 
Encheleus, Autarieus, Dardanus, Maedus, 
Taulas, and Perrhaebus, also daughters, 
Partho, Daortho, Dassaro, and others, from 
whom sprang the T aulantii, the Perrhaebi, and 
Enchelees. the Autarienses, the Dardani, the 
Partheni, the Dassaretii, and the Darsii. 
Autarieus had a son Pannonius or Paeon, and 
the latter had sons, Scordiscus and Triballus, 
from whom also nations bearing similar names 
were derived. But I will leave these matters to 

(Applan, The lllyrian Wars, 2) 

Daorsi Autariates 

■ ••.•••rO'/ ^ 


Docleatae Soirtones 

: ■> Parthini 

45 The lllyrian tribes and one solution to their origins by the second century ad historian 



offering an unbroken cultural succession from the Early Bronze Age to well into 
the Iron Age and the historical appearance of Illyrian tribes; this is seen 
especially in the great multiperiod tumulus cemetery of Glasinac. In Bosnia, 
too, there is frequent mention of this pre-Bronze Age culture break generated 
by intruding pastoralists who buried their dead under tumuli. 

Although archaeologists identify this intrusion with the influx of Indo- 
European peoples, they and their colleagues in the field of linguistics are careful 
not to identify the invaders with the actual Illyrians. Rather, they view Illyrian 
origins, like Thracian, as an extremely complicated process involving an 
intrusive Indo-European-speaking component mixing with local populations 
and eventually, by the Iron Age, giving rise to linguistically related tribal 
groups which we must somewhat uncomfortably label as Illyrians. 

The Slavs 

The largest group of Indo-European-speaking peoples of Eastern Europe is the 
Slavs. Today there are about 430 million speakers of Slavic languages and with 
Russian as the lingua franca of both the European and Asiatic portions of the 
USSR, it is also one of the most rapidly expanding Indo-European groups. 
This expansion, as we will soon see, has been relatively recent when compared 
with most other Indo-European groups. 

According to both historical tradition and the available written evidence the 
earliest Slavic texts date only to about the ninth century ad when the 
missionaries Constantine (Cyril) and Methodius devised the Cyrillic alphabet 
and translated portions of the Bible and the Eastern Orthodox liturgy into what 
we now term Old Church Slavonic. This is by no means our first historical 
acquaintance with the Slavs since the Sclavini, Antes and Veneti - all probable 
Slavic ethnic groups - were known several centuries or more before the earliest 
written Slavic texts. It was historians of the eastern Roman empire who 
recorded the explosion of Slavs into the Byzantine world from the sixth century 
ad onwards in a virtual litany of raids and invasions that carried the Slavs into 
the Balkans and Greece. Historians such as Procopius and Jordanes, writing 

Old Church Slavonic: 



















































da se 











da se 













46 The opening line of the Lord's Prayer in Old Church Slavonic plus representative 
languages from the major modern Slavic groups indicates the type of close similarities one expects 
from a language that has differentiated in the not-too-distant past. The Slavic word for 'heaven 
nebesa, nebo preserves the Proto-Indo-European word for "cloud, mist, heaven' seen in Hittite 
nepis, Sanskrit nabhas, Greek nephos, Latin nebula, Old English nifol, and Irish neamh. 



47 The distribution of the 
Shi vie languages. The eastern 
group is indicated with 
diagonal hatc hing, the western 
with vertical and the southern 
group with horizontal 

about the sixth to seventh centuries ad, locate the Slavic tribes of the Sclavini 
and Antes north of the Danube in a band stretching from the upper Vistula to 
the Dnieper. The earlier attested Veneti, known from the first to the second 
centuries ad, are more difficult to place, although this same general region 
would not be too far off most attempts to locate them more precisely. 

Both linguists and archaeologists concerned with the origin of the Slavs see 
their borders as extremely dynamic from the fifth to the tenth centuries AD. 
This period is not only regarded as the primary time of Slavic expansions but 
also as the terminal period of proto- or Common Slavic. By this is meant that for 
the period ad 400-900 linguists recognize the collapse of Common Slavic and 
its fission into the different modern Slavic languages. Expansions to the east 
and northeast from ad 500 to 1000, for example, carried Slavic speakers into 
territories previously occupied by Baits and Finns. Today these eastern Slavs 
are represented by the Russians, Byelorussians (White Russians) and 
Ukrainians. The southern expansion across the Danube into the Byzantine 
empire, and subsequent divisions, account for the southern Slavs - Bulgarians, 
Macedonians, Serbs, Croats and Slovenians. In the west emerged the Poles, 
Czechs and Slovaks. This great fission of the Common Slavic language 
occurred very broadly at the same time as Latin was disintegrating into the 
various Romance languages, albeit under quite different social circumstances. 

Linguistic evidence indicates that before the collapse of Common Slavic, 
that is, before the fifth century ad, the Slavs had been subjected to strong 
linguistic influences, primarily seen in loan words, from Germanic-(Gothic) 
and Iranian-(Sarmatian) speaking peoples. It is from the Sarmatians (or the 
Scythians), for example, that Slavicists derive the Common Slavic words for 
4 god\ 'holy' and 'paradise' plus perhaps several score more terms. Even the 
names of the major rivers of the European steppe - the Don, Dnieper and 
Dniester -are all of Iranian origin. In addition, before the collapse of Common 



Slavic there had been an extremely complicated (and controversial) relation- 
ship with speakers of the Baltic languages. This relationship has been seen by 
some as an intensely close genetic connection such that a common Balto-Slavic 
emerged out of late Proto-Indo-European (much like Indo-Iranian) before 
further dividing into Baltic and Slavic. Others prefer to see the similarities 
between them as contact relations betw een two adjacent languages that evolved 
independently from late Proto-Indo-European without an intermediary Balto- 
Slavic stage. It would be sheer folly to wade into the linguistic morass of Baltic- 
Slavic relations in a work such as this, and we will content ourselves with the 
minimal conclusions that both Baits and Slavs, at some time prior to the fifth 
century ad, lived in close proximity with one another, and indeed they would 
appear to have done so for much of their existence. The linguistic evidence 
positions the Slavs to the east or southeast of the Germans, south of the Baits 
and west of the Iranians. 

The problem of the earlier location of the Slavs prior to their first entry into 
the historical records of the Byzantine state is a perennial occupation of Slavic 
linguists and archaeologists. While there is no total consensus, one can extract 
areas of relatively general agreement. Certainly it is not controversial to assert 
that about ad 500 the Slavs occupied some or all of the territory stretching in a 
wide band from the Elbe, Oder or upper Vistula on the west to at least the 
middle Dnieper on the east. Controversy only really begins when we try to 
identify the Slavs more specifically within this general region or at an earlier 
date. The Ukrainian archaeologist, Vladimir Baran, for example, initiates his 
study of the proto-Slavs w ith the historically attested peoples of the sixth to 
seventh centuries ad and assigns them to the Prague-Penkov-Kolochina 
complex that inhabited a very broad area from the Elbe on the west to beyond 
the Dnieper on the east. Throughout this area we find generally similiar 
ceramics coupled with semi-subterranean rectangular houses, hearths and 
cremation burials in urns. The general uniformity of the archaeological groups 

48, 4g The distribution of Iranian river names in Eastern Europe. These include the major 
rivers whose names are built from the Iranian word danu- 'river, whence the Don, the Dnieper 
(from *danu apara 'river to the rear ) and Dniester f*danu nazdya- 'river to the front' ). The 
same Indo-European root underlies Celtic Danuvius, Danube. On the right is the Prague- 
Penkov-Kolochina complex. 



accords well with both our historical sources and our expectations based on the 
later appearance of Slavic peoples beyond the Byzantine frontier. 

Attempts to push back beyond the fifth century ad and follow a trail of direct 
cultural continuity must somehow penetrate through the 'noise' produced by 
the Chernyakovo complex of the second to the fifth centuries ad. This culture, 
evidenced by more than a thousand sites, embraced most of the northwest 
Pontic region from the Danube northeast to the Seym. Its existence runs 
concurrent with the intrusions of the Goths from the northw est and Sarmatians 
from the east, all into an area with residual populations of Hellenized Scythians, 
Getae, Dacians and probably Slavs. Some argue that the Chernyakovo culture 
may be directly equated with the Goths while others deny the Goths any 
archaeological visibility and insist on interpreting this complex as a sort of 
melange created among a number of very different local groups. A hint of the 
controversy can readily be seen when we recall that Marija Gimbutas regards 
this period as the time when Goths superimposed themselves politically on the 
Slavs and consequently passed on to Common Slavic a number of Germanic 
loan words such as 'bread', 'house', and 'stable'. Joachim Werner, on the other 
hand, regards the Chernyakovo culture as exclusively Germanic and places the 
Slavs further north in the forest, unable to push southwards until the Huns 
eliminated Gothic settlement and power in the forest steppe in the fourth 
century ad, while Valentin Sedov sees the northern region of the Chernyakovo 
culture as the area w here Iranians and Slavs achieved their sy mbiosis. Basically, 
the Chernyakovo complex appears to have embraced a variety of different 
ethno-linguistic groups and we would assign only portions of it to Slav, Iranian, 
Goth or any another group. 

Agreement is still more difficult to find when one retreats back further in 
time to the period of the Zarubinets-Przewor cultures of the second century 
BC-second century ad. The more easterly culture, the Zarubinets, is seen by 
many today as an unquestioned precursor to the later historically attested Slavic 

50, 5/ The Chernyakovo cultural region which appears to have included Slavs, Goths and 
Iranian-speaking tribes. On the right is the Zarubinets-Przewor complex. 



cultures. It is the more westerly situated Przewor culture of the Elbe-Vistula 
region, widely regarded as related to the Zarubinets, that provokes the most 
controversy. Its inclusion within the Slavic 'homeland' is in accord with the 
Polish hypothesis w hich sees cultural continuity in the Elbe-west Baltic region 
from the Bronze Age until the historical emergence of the Slavs. This 
continuity sufficed to justify the theory that the Slavs originated in this very 
region before expanding south and eastwards. In opposition to this theory is the 
frequently cited opinion of Germanicists that the Przewor region is more 
arguably within an early Germanic-speaking territory either of, or immediately 
adjacent to, the Elbe-Germanic tribes. Battle-lines drawn along a prehistoric 
border, one scholar jokingly observed that 'German scholars would like to 
drown all the Slavs in the Pripet swamps, and Slavic scholars all the Germans in 
the Dollart'. A resolution of these conflicting theories generally rests on 
inconclusive arguments for regional cultural continuity, the geographical 
deadreckoning generated from ambiguous historical sources such as Tacitus or 
Ptolemy, and Old Slavic river names. 

There is a wide acceptance among both linguists and archaeologists that the 
study of river names can provide an important source of prehistoric 
information. The work of Oleg Trubachev, for example, provides a relatively 
well-defined zone of rivers retaining archaic Slavic names and neighboured by 
non-Slavic hydronymic systems. This archaic Slavic hydronomy is confined 
primarily to the region stretching from the upper Vistula basin to the middle 
Dnieper. Many archaeologists have accepted this as an important confirmation 
of their archaeological theories, although one cannot escape remarking that this 
common Slavic hydronomy is not an especially well-dated phenomenon and 
archaeologists can be rather cavalier in how they utilize it. Some, for example, 
apply the hydronymic map to the archaeology of the first centuries AD and see 
confirmation that the southeastern Przewor, part of the Chernyakovo and the 
Zarubinets cultures are all included as a Proto-Slavic homeland. Others see a 

92. S3 The area of old Slavic river names (dotted line) plotted against the distribution of the 
Chernoles culture. On the right is the Komarov c ulture and its eastern neighbour^ the Trzciniec 
culture; the latter has been variously assigned tit both the earliest Slavs and the earliest Baits. 



far better 'fit' by overlaying the river-name maps on the still older Chernoles 
culture of the period 750-200 BC. This area coincides with the territory 
attributed by Herodotus to Scythian-Farmers and a reasonable case can be 
made for seeing these Scythian-Farmers as Iron Age Slavs. 

The earliest that Slavicists are generally willing to push the concept of a 
linguistically differentiated Proto-Slavic is 2000-1500 bc. The most widely 
accepted archaeological representative of this earliest Slavic period is the 
Komarov complex which dates to about 1500 BC and which occupies the region 
of the middle Dnieper to the upper Vistula. Again we confront the problem of a 
more westerly extension since both the Polish hypothesis as well as a number of 
Soviet scholars would also include the Trzciniec culture to its northwest as a 
related and hence Proto-Slavic culture. The Komarov culture itself is known 
especially from its burials which are primarily inhumation in a timber or stone- 
covered grave with a low tumulus. 

The general course of investigations into Slavic origins may be laden with 
controversy but these differing opinions are normally confined to the western 
limits of early Slavic territory. It is difficult to deny that there existed a 
geographical centre weighted between the Vistula and Dnieper which is most 
commonly agreed to be Proto-Slavic and which appears to display a continuity 
of cultural development from about 1500 bc (or earlier) to the historical 
appearance of the earliest Slavic peoples. To derive the Slavs exclusively from a 
more westerly area such as the Elbe-Vistula region requires far less economy of 
explanation, not to speak of movement; moreover, the Iranian loans into 
Common Slavic make the case for a more easterly home for the early Slavs more 
attractive. It must be admitted that, throughout all of these arguments, we find 
ourselves engaging in the archaeologists' easiest pursuit - the demonstration of 
relative continuity and absence of intrusion. A long geographical stasis for the 
Slavs, however, is probably the model that would be most readily accepted by 
linguists who see in the Slavic language group little reason to assume that they 
have moved much since their development from Proto-Indo-European. 
Whether this can be employed as an anchor for yet earlier Indo-Europeans is, of 
course, another matter which we will have to investigate later. 

The Baits 

The course of expansion that carried the Slavic languages over much of Eastern 
Europe was also responsible for greatly reducing the area occupied by Baltic 
speakers. Today there are an estimated six million speakers of the two surviving 
east Baltic languages, Lithuanian and Latvian. The major representative of the 
western Baltic languages, Old Prussian, became extinct by about ad 1700. The 
pressure of Slavic expansions from the south and Germanic from the west has 
reduced the original Baltic-speaking territory to an estimated one-sixth of its 
previous area. 

The earliest Baltic texts appear quite late when compared with most other 
Indo-European languages. In the sixteenth century we first encounter written 



examples of both the Old Prussian and Lithuanian languages generally 
emerging in the form of religious literature such as Lutheran catechisms. The 
texts, as indeed the modern Lithuanian language today, has always attracted the 
attention of linguists since, despite their recent date, they appear remarkedly 
archaic in terms of Indo-European linguistics. To take a familiar example, the 
Lithuanian proverb l God gave teeth; God will give bread' displays an almost 
incredible similarity to its translation into the much older Latin and Sanskrit: 

Lithuanian Dievas dave dantis; Dievas duos duonos 
Sanskrit Devas adadat datas; Devas dat dhanas 
Latin Deus dedit dentes; Deus dabit panem 

Because of this transparent conservatism, many linguists hold that the Baltic 
languages, like their Slavic neighbours, have probably moved but little since 
late Indo-European times. 

Although the earliest written texts date to the sixteenth century, this is hardly 
a terminal point in our quest for Baltic origins. By this time much of Eastern 
Europe had already been encapsulated in a vast Lithuanian state, dating from 
the fourteenth century, w hich had witnessed the greatest expansion of Baltic 
political power in historical times. The Baits are also well attested in tribal and 
personal names as they confronted the Teutonic Order in the course of the 
thirteenth century. Before that we find them, especially in the form of the 
coastal Curonians, harrying the Baltic in the tenth and eleventh centuries, and 
they are mentioned by Scandinav ians and the Anglo-Saxon Wulfstan from the 
seventh to the ninth centuries ad. Throughout all of these periods we find them 
in their historical homes centred on the Baltic Sea from the Vistula north to at 
least the Daugav a. By this time, however, they have already undergone massive 
retraction in the east due to the pressure of East Slav ic expansion, attested by 
finds typical of the Prague and other Slavic archaeological complexes. 



Menuo sauluze vede 

The Moon leads (home) the Sun, 

Pirma pavasareli 

In the first of spring, 

Sauluze anksti keles 

The Sun rose early, 

Menuzis atsiskyre 

The Moon left her. 

Menuo viens vaikstinejo 

The Moon alone wandered, 

Ausrine pamylejo 

With the Morning Star he fell in love. 

Perkuns, didziai supykes 

Perkunas, very angry, 

Ji kardu perdalijo 

With his sword he cut (him) to pieces. 

55 An excerpt from a Lithuanian folksong which still reflects the pagan traditions of the 
ancient Baits. All of the celestial words here have good Indo-European origins: menuo 'moon is 
cognate with a whole series of Indo-European words meaning moon and month, e.g., Sanskrit 
mas, Greek men, Latin mensis, Tocharian man and, of course, English moon. Saule 'sun' goes 
with Sanskrit suvar, Latin sol, Gothic sauil, etc., while ausrine is u diminutive o/'ausra 'dawn 
and belongs with Sanskrit usas, Greek eos, Latin aurora, and English Easter all of which are 
similarly personified or deified. The verb vede is cognate with English wed and other Indo- 
European verbs derived from *wedh 'to lead home, to marry (from the groom's point of view)' 
which suggests that an early Indo-European bride moved to the home of her husband or his 
family. Perkunas is the archetypal thunder-god; his name is cognate with the Slavic god Perun 
and (he same root furnishes us with the name of Fjbrgyn, the mother of the Norse thunder-god 

We can recede further into Baltic history via classical authors such as 
Ptolemy and Tacitus who list the names of Baltic tribes including the famed 
gatherers of amber, the Aisti - perhaps to be equated with western Baits - and 
tribes such as the Soudinoi and Galindai who emerge in the fourteenth century 
as the names of Prussian tribes. The only available historical reference beyond 
this period is Herodotus who speaks of a people known as the Neuri. Herodotus 
tells us little of the Neuri other than how they were driven from their country in 
the sixth century bc because of an enormous plague of snakes, and that for one 
or two days a year they transformed themselves into werewolves (a story that 
even Herodotus himself dismissed). Interestingly enough, the zaltys, the 
Lithuanian green snake, occupied an extraordinarily important role in pagan 
Li thuanian belief as well as folklore. We read how each household maintained 
its own snake, and contemporary accounts of the forced Christianization of the 
Baits by their German neighbours report how the Germans seized the snakes 
against great protest and burnt them in public bonfires in the middle of the 
Baltic villages. 

Herodotus locates the Neuri north of the Scythian Farmers, the probable 
designation of the Slavs, and separated from the Scythian world by a lake, 
possibly the Pripet marshes. This locates the Baits in the territory assigned by 
archaeologists to the Milograd culture which occupied the northern Dnieper 
Basin, and it falls well within the Old Baltic hydronymic system. 

Like Slavic river names, there has been a long tradition of research into the 
distribution of Baltic river names which clearly indicates that the Baltic 
linguistic area once included regions far to both the east and south of the coastal 
Baits, whom we encounter in most historical records from the Middle Ages 
onwards. The system is generally complementary to the Slavic river names, 
occupying the territory immediately to the north of the Slavs and bordered on 



the northeast by Finno-Ugric names. We naturally confront the same problem 
of their chronology which we encountered with Slavic river names. They must 
certainly predate the Slavic expansions of the fifth century ad and if truly 
coordinate with the Slavic area designated by Herodotus and the Chernoles 
culture, then we may place them to 500 bc. Wolfgang P. Schmid, however, 
would push them back into the greatest depths of Proto-Indo-European. Little 
is to be gained by overplaying the hydronymic card and asserting for one 
language that which cannot be demonstrated for any other. The minimal view 
would see the Baits, during the first millennium BC, occupying the area from 
west of the Vistula's mouth east to Moscow and the upper Volga (which itself 
carries a Baltic name), and south almost as far as Kiev. It may have been at this 
time or possibly earlier that a large number of Baltic words, especially those 
concerned with the technology of agriculture and stockbreeding, entered the 
Finnish language. 

To recede any further into prehistory requires us to assume, as with the 
Slavs, direct continuity of culture and hence no evidence for movement of 
peoples within this region. This, so the argument goes, would carry us back to 
Middle Bronze Age cultures of about 1 500 bc, including the Trzciniec culture 
of Poland which some have included as a Slavic culture. Although continuity 
can be argued to still earlier periods, it is useful to halt here, as we have done 
with other Indo-European groups, since we can only compound uncertainties 
by extrapolating further back into prehistory. 

The Germans 

The modern Germanic languages are traditionally divided into two major 
groups. The larger West Germanic group comprises English which is spoken 
by over 425 million speakers; German, spoken by approximately 120 million; 
and Dutch and its derivative Afrikaans which are spoken by about 30 million 
people. The smaller group consists of the North Germanic or Scandinavian 
languages which today embrace 20 million speakers of Danish, Swedish, 
Norwegian and Icelandic. Although many of these languages, especially 
English, German and Icelandic, offer abundant texts from the so-called Dark 
Ages and medieval period, our earliest substantial texts are in the extinct East 
Germanic Gothic language. The Goths had migrated from the north into the 
Black Sea region where they ruled until the arrival of the Huns in the fourth 
century ad who pushed most of the Gothic tribes westward into the Balkans. It 
was here that Wulfilas, Bishop of the West Goths, created a Gothic alphabet 
(primarily derived from the Greek) and translated portions of the Bible into the 
Gothic language. Gothic was still spoken up into remarkably recent times, 
surviving in the Crimea as late as the sixteenth century. 

Other than Gothic texts we also have the evidence of the runic inscriptions 
which appear to have been loosely derived from one of the north Italian or 
Etruscan alphabets of the first century ad. Runic inscriptions were employed 
widely over northern Europe from about AD 150 to 900, and still earlier is the 



solitary evidence of the Negau helmet, discovered in Yugoslavia and bearing 
the inscription harixasti teiva which is usually translated 'to Army-Guest (or 
Host), the god', putatively a dedicatory inscription to a Germanic god. The 
helmet is inscribed in a north Italian script and has been variously dated from 
about the seventh to the second century bc. 

Other than direct textual evidence of the Germanic languages, we have the 
writings of classical historians, the most important of w horn is Tacitus who 
described the location and culture of the early Germans in his Germania. 
Tacitus located most of the Germanic tribes about ad 100 in a region bordered 
on the west by the Rhine, to the south by the Main and on the east by the Oder. 
Earlier sources, such as Caesar in the first century bc, are somewhat less reliable 
but again place the Germans east of the Rhine. The earliest historical source, 
Pytheas, is generally understood to have located the Germanic tribe of the 
Teutones in present-day Denmark and the Gutones possibly in northern 

Both the historical and inscriptional evidence indicate that the earliest 
Germanic-speaking peoples were largely distributed in the area of northern 
Germany and southern Scandinavia. To confine the Germans more precisely 
within this area we must have recourse again to the study of river names and 
early tribal names, especially among peoples lying immediately east of the 
Rhine. A few linguists suggest that the territory between the Oise and Aller may 
have been occupied by a linguistic group which was neither Celtic nor 
Germanic but which has been termed the Nordwestblock. 11 Although the 
existence of such a grouping in the first centuries bc is not universally agreed, 
the hypothesis at least poses a salutary reminder that some historically 
anonymous groups may have survived to the dawn of historical records and we 

56 The distribution of the 
Germanic languages. The 
northern group is indicated 
by vertical hatching, the 
western group with horizontal. 
The eastern Germanic group, 
attested by Gothic, is extinct. 



The Germanic Languages 

Gothic (fourth century) 

Jah hairdjos wesun in thamma samin landa thairhwakandans jah 
witandans wahtwom nahts ufaro hairdai seinai Ith aggilus fraujins 
anaqam ins jah wulthus fraujins biskain ins, jah ohtedun agisa 

Old English (tenth/eleventh century) 

& hyrdas waeron on tham ylcan rice waciende. & niht-waeccan 
healdende ofer heora heorda. Tha stod drihtnes engel with hig & 
godes beorhtnes him ymbe-scean. & hi him mycelum ege adredon 

Middle English 

And schepherdis weren in the same cuntre. wakinge and>iepinge the 
watchis of the nyzt on her f lok. And loo! The aungel of the Lord stood by 
sydis hem, and the clerenesse of God schynedet aboute hem; and thei 
dredden with greet drede 

Low German (fifteenth century) 

Unde de herden weren in der suluen iegenode wakende. Unde helden 
de wake auer ere schape. Unde seet de engel des heren stunt by en 
unde de clarheit godes ummevench se unde se vruchteden sick myt 
groten vruchten. 

High German (sixteenth century) 

Und es woren Hirten in derselbigen Gegend auf dem Felde bei den 
Hurden, die huteten des Nachts ihrer Herde. Und siehe, des Herrn 
Engel trat zu ihnen, und die Klarheit des Herrn leuchtete urn sie; und 
sie furchteten sich. 


I samma nejd voro da nagra herdar ute pa marken och hollo vakt om 
natten over sin hjord. Da stod en Herrens angel framfor dem, och 
Herrens harlighet kringstralade dem; och de blevo mycket 

And shepherds were in that same land abiding and keeping watch by 
night over their flocks. But the angel of the lord approached them and 
the glory of the lord shone about them, and they feared greatly. 

57 A brief excerpt from the Gothic Bible compared with other Germanic groups. The hairda 
'flock 1 is to be compared with English herd as well as Sanskrit sardha-, Lithuanian kerdzius 
and Middle Welsh cordd 'troop'. As the word nahts 'night' is part of very basic vocabulary, we 
find it very well attested across the Indo-European languages, e.g., Htttite nekut, Sanskrit nak, 
Greek nyks, Albanian nate, Latin nox, Old Irish in-nocht, Lithuanian naktis, Old Church 
Slavonic nosti. The word for [fear agis goes with Greek akhos and Old Irish agor 'fear. 

would be ill-advised to imagine that all of Europe or Asia was occupied by only 
those branches of the Indo-European languages recorded in history. 

The area demarcated by the historical and linguistic evidence back to the first 
century BC has provided archaeologists with an independent sense of security in 
their identification of the earliest Germanic region. Continuity of settlement 
and culture can be clearly observed from the initial historical emergence of the 
Germanic tribes in the first centuries ad back to the Iron Age Jastorf culture. 
This culture, appearing about the fifth century BC, was the dominant Iron Age 



culture of Northern Europe and the continuity of its settlements, cemeteries, 
and distribution accords well with the historical locations of the earliest-known 
Germanic tribes. Consequently, the Jastorf culture and probably the neigh- 
bouring Harpstedt culture have provided Germanicists with a generally 
agreed-upon Germanic homeland. This also finds reinforcement from the 
linguists who suggest that those sound changes that transformed a late Indo- 
European dialect into Proto-Germanic probably occurred about 500 bc. 

There is a strong temptation to push the Proto-Germanic presence in this 
same region back further in time. Certainly, no major body of archaeologists 
would argue that the Jastorf culture was anything other than a direct 
descendant from the Later Bronze Age of the same area. This can be 
demonstrated by the continuity of settlement and cemetery practices which are 
as much in evidence here as between the Jastorf culture and the historical 
period, but we cannot really penetrate beyond this and still hope to retain the 
name Proto-Germanic in a linguistically meaningful sense. 

We can see, then, that the Jastorf culture w as Proto-Germanic by common 
agreement. What preceded it may also have been Proto-Germanic or perhaps 
late western Indo-European, or some other state of the evolution of the Indo- 
European languages for which we have no precise name. We will have good 
reason to return to this problem at the end of this chapter. 


While many may carp with Metternich's dismissal of Italy as not a nation, but 
only 4 a geographical expression', this is a particularly apt description for the 
Indo-Europeanist. By the end of the eighth century bc, Greek colonists in 
southern Italy had introduced the alphabet which subsequently spread 
throughout the Iron Age cultures of the Italian peninsula. The resulting 
inscriptions have opportunely preserved the evidence for a variety of different 



languages that were, for the most part, soon to become extinct. A linguistic map 
of Italy set to a notional date of 500 bc gives some indication of the linguistic 
complexity of Italy prior to the expansion of Latin by the Romans. It is for this 
reason that we must now abandon our usual framew ork of presenting an Indo- 
European group under its ethno-linguistic heading, for we will have to confront 
all of the linguistic evidence from Italy at one time. 

In discussing the linguistic history of Italy it is best to begin with the 
evidence for non-Indo-European languages. By far the most important 
example is Etruscan which provides us w ith over 10,000 inscriptions and some 
short texts in a language which the overwhelming majority of linguists have 
concluded is not Indo-European and not demonstrably related to any other 
language except for some inscriptional evidence on the island of Lemnos in the 
east Mediterranean. This raises the entire problem of Etruscan origins which 
has filled volumes every bit as large as this one and is as heatedly debated as any 
of the problems concerning Indo-European origins. There is no easy solution, 
since the evidence is extremely self-contradictory. Nevertheless, the present 
tendency in Etruscan research is to adopt the most economical thesis: the 
Etruscans were a non-Indo-European people native to Italy who adopted many 
items and styles of east Mediterranean provenience by way of trade. The 
similarity between Etruscan and the Lemnian inscriptions must be acknow- 
ledged and is admittedly difficult to explain. One thesis sees both Etruscan and 
Lemnian as remnants of a continuum of non-Indo-European 'Mediterranean' 
languages which spanned the eastern and central Mediterranean before the 
intrusion of Indo-European speakers. But the similarity between Etruscan and 
Lemnian is, I believe, too great to be explained by anything other than a more 


59 The distribution of the 
ma 1 or languages of Iron Age 




















*k w etwores 




*penk w e 





















6o The non-Indo-European nature of Etruscan is clearly seen in this comparison between it, 
Latin and Proto-Indo-European. 

direct and immediate historical connection, possibly involving a visit to 
Lemnos by Etruscan traders. 

Perhaps one of the reasons that it is slightly easier to accept Etruscan as a 
native language of Italy is that there is other evidence for non-Indo-European 
languages in the region. A large body of linguistic opinion supports the 
argument that both place names, particularly in the western Alpine region and 
in Sardinia, and many of the words in Latin and the Romance languages 
unanalysable from the standpoint of Indo-European, derive from a pre-Indo- 
European substrate. This evidence is, of course, very similar to that which we 
have seen employed to substantiate a pre-Greek population and it must be 
admitted that only the Etruscan language offers us conclusive textual evidence. 
Some have suggested that Ligurian, a language attested to the north of the 
Etruscans in a few glosses and local names, was also non-Indo-European 
(although heavily influenced by the Celts), but the linguistic evidence here is far 
too sparse to draw any firm conclusions. Similarly, in the eastern Alps where 
Raetic is also slightly attested, there have been claims that this language 
possesses distinctly non-Indo-European elements, for example, Raetic tinake, 
Etruscan zitutke, but here again the inscriptional evidence is reallv verv meagre. 
Nevertheless, when we combine all this evidence with Etruscan, the non-Indo- 
European place names and elements in Latin, and the proximity of at least 
northern Italy to conclusively non-Indo-European languages in southern 
France and Iberia, we have little doubt that the Indo-European languages were 
intrusive into Italy and that they were superimposed on a variety of non-Indo- 
European populations. 

The most famous of the Indo-European languages in Italy is naturally Latin 
whose spread coincided with the expansion of the power of the Roman state. It 
is a tribute to the persistence of the Roman conquest that, out of the vernacular 
language spoken in its territories at the height of the Roman empire, there 
emerged the Romance languages of French, Spanish, Portuguese, Italian, 
Romanian and others, with approximately 550 million native speakers today, 
rendering Italic the second largest group of Indo-European speakers. But at the 
time we have drawn up our linguistic map, Latin would appear to have been 
confined to the territory about Rome with its nearest linguistic neighbour seen 


in the dialect of Faliscan immediately to its north. Far more controversial are 
attempts to relate Latin closely to Siculan, a language spoken in eastern Sicily 
and attested by only three inscriptions and some glosses. 

Down the spine of Italy ran the major group of Osco-Umbrian. Oscan itself 
was the language of the Samnites and it probably did not become extinct until 
the first centuries AD if graffiti on the walls of Pompeii can be any guide. There 
are about 200 inscriptions plus the usual glosses and personal names. Umbrian 
is better attested with the Iguvium tablets, religious texts inscribed on bronze 
which date back to about 200 BC. The differences between Latin and Osco- 
Umbrian are as obvious as their similarities and linguists are unsettled as to 
whether a common Proto-Italic dialect was carried into Italy from whence the 
two major groups developed, or whether they represent independent develop- 
ments, perhaps outside of Italy, which have gained similarities through long- 
standing contacts. 

The major language of southeastern Italy, apart from that of the Greek 
colonists, was Messapic, which is known from about 260 inscriptions dating to 
approximately the sixth to the first centuries bc. A combination of ancient 
historical testimony tracing Messapic tribes to Illyria coupled with archaeo- 
logical evidence for cross-Adriatic connections in ceramics and metalwork have 
prompted linguists to link Messapic with Illyrian. The hard linguistic evidence 
is minimal since there are no Illyrian inscriptions and the link must be based on 
personal and place names. 18 Consensus does support such a link but not without 
hesitation; a minority is more cautious and does not regard Messapic-Illyrian 
links as even discussable. 

To the north of Messapic is Picene or East Italic which may comprise two 
different languages under the same name. Some of the Picene inscriptions 


este persklum aves anzeriates enetu 
pernaies pusnaes preveres treplanes 
iuve krapuvi tre buf fetu arvia ustentu 
vatuva ferine feitu heris vinu heri puni 
ukriper fisiu tutaper ikuvina feitu sevum 
kutef pesnimu arepes arves 


hanc caerimoniam avibus observatis 

anticis posticis. ante portam Trebulanam 

lovi Grabovio tres boves facito, grana ostendito, 

victimas in tabulato facitio, sive vino sive potione, 

pro arce Fisia, pro urbe Iguvina facito. 

formulam dare precator tostis granis. 

Commence this ceremony by observing the birds, 

those from in front, and those from behind Before the Trebulam Gate 

sacrifice three oxen to Jupiter Grabovius. Present grain offerings, 

place the ribs on a tray, sacrifice with wine or with mead, 

for the Fisian Mount, for the state of Iguvium. 

Pray each (portion) in a murmur with (offerings) of fat and grain. 

61 An excerpt from the beginning of the Iguvium tablets with Latin and English translation. 





'Here, Jupiter! Thotoria Marta gave to the city of Basta her land in [the 
locality] of Darantoa . . .' ('Consensus' translation) 

Here, Jupiter! Under the *teutor-ship of Amartapidius, for the 
inhabitants of Basta this law. They passed it (the law) in the Council of 
Elders in Basta . . ." (Otto Haas translation) 

'One should heed these things! I, Teutoria Marta, purchase in the city 
of Basta the wine fields in Darantua . . .' (M E. Huld translation) 

62 This inscription from Vaste is one of the longest texts in the Messapic language and 
illustrates the problems in translating such marginally attested languages. The original 
inscription runs for eight lines and there are no breaks between words. These must be determined 
by isolating out repeated words or word-endings from which very conflicting translations may be 

extend back to about the seventh century bc and are among our earliest written 
evidence in Italy. Unfortunately, the inscriptions, though easily deciphered, 
are not easily translated. The southern inscriptions are at least transparently 
Indo-European for example, matereif patereif= Latin matribus patribus 'to the 
mothers and fathers', while the north Picene inscriptions pose far more 
problems. These have been judged intuitively by some linguists to look like 
Indo-European (even though we cannot translate a single word confidently!) 
while others see them as evidence of the non-Indo-European natives. Those 
who accept an Indo-European identification frequently derive them, like 
Messapic, from the east Adriatic. 

Finally, in the Veneto to the northeast we have Venetic, the language of the 
Iron Age Este culture which is at least unqualifiably Indo-European. The over 
200 inscriptions (though none over ten words long) were written from the sixth 
to the first centuries bc. In addition, there is toponymic evidence which here 
relates the territory of the Veneti to the tribe of the Liburni of Adriatic 
Yugoslavia. Venetic is sufficiently well known to argue about its linguistic 
affinities with other Indo-European languages. Similarities with Italic and 
Germanic have all been suggested, especially with the former, but some 
linguists regard Venetic as an independent Indo-European subgroup which 
shares some similarities with other West European languages but not enough to 
link it closely with any particular one. 

When reviewing the origins of the Indo-European peoples of Italy we may 
accept that they were intrusive and superimposed on a non-Indo-European 
substrate. Since we find the evidence for non-Indo-European languages 
increases as we move westwards either through Italy or along the Mediterra- 
nean coast, the most probable direction for Indo-European movements is either 
from north of the Alps or across the Adriatic from the east. This becomes 
particularly plausible when we consider that by the ninth century bc there had 



North Picene 

mimnis erut gaarestades 

rotnem uvlin parten us 

polem isatron tet 

sut trat nesi krus 

6 j The famous Novilara inscription is the longest 
example of the north Picene language. Scholars are 
divided on whether the language is Indo-European or 
not but are generally agreed that there is not a single 
word in this inscription that can confidently he 


tenag trut ipiem rotnes 
lutuis thalu isperton vul 
tes rotem teu aiten tasur 
soter merpon kalatne 
nis vilatos paten arn 
uis balestenag ands et 
sut i akut treten teletau 
nem polem tisu sotris eus. 

developed three major archaeological provinces along Adriatic Italy which 
coincide well with three of our major linguistic groups. In the Veneto we find 
the Este culture first emerging at 900 bc, while both a Picene archaeological 
region and the Late Bronze Age cultures of Apulia and Basilicata provide our 
second and third archaeological units. All three of these provinces show 
longstanding and intense relationships with the east Adriatic, especially in 
metalwork, such that regular movements of people across the Adriatic Sea or 
along the northeast coast of Italy have frequently been postulated. These 
Balkan currents are relatively constant from at least the thirteenth century bc 
onwards, and they prevent us from chasing the illusive and oversimplistic 
model of a single invasion to explain the Picenes and the Messapi. Rather, there 
may have been long-term movements of Indo-European speakers into Adriatic 
Italy who, by dominating the larger and richer coastal sites, were in an optimum 
position to spread their languages into the smaller and poorer interior. These 
movements from the east would appear to define the most recent layer of Indo- 
European immigrants except, of course, for the fifteenth-century Albanian 
refugees who also crossed the Adriatic to settle in southern Italy and Sicily 
where today they still retain their own language. 

Another putative intrusion is associated with the rapid spread of the Proto- 
Villanovan culture (or Pianello-Timmari horizon) which emerged over most of 
Italy about 1 100-900 bc. Although it has long been considered a prime vector 
of the Indo-European languages, it must be admitted that it seems to provide us 
with precisely the type of evidence for which we are not looking in order to 
explain the linguistic diversity of Italy, since it displays a remarkable cultural 
uniformity over almost the entire peninsula. It is in this period that we see the 
widespread appearance of cremation cemeteries employing biconical urns, and 
a wide variety of Central European metalwork that includes fibulae, razors, 
pins, swords, and sheet bronze work such as buckets, helmets and armour. The 
introduction of cremation with urns, and the metalwork, have been attributed 
in the past to Urnfield invaders from the north of the Alps who penetrated 
northern Italy to produce a major upheaval in culture. Present archaeological 
models dismiss schemes of massive Urnfield invasions as grossly simplistic 
relicts of all-too-traditional archaeological thinking. Nevertheless, attributing 



the similarities between Italy and Central Europe to vaguely defined models of 
persistent cultural contacts or orientations still leaves room for some element of 
folk-movement, even if nowhere on the scale of those who dreamed of Late 
Bronze Age warriors bursting through the Alpine passes. The results of such 
movements, however, become linguistically obscure when we recall that the 
same cultural phenomenon which underlies areas of Indo-European speakers 
such as the Osco-Umbrians, also provides the immediate foundation of the 
Villanovan culture of Etruria which we recognize as the earliest archaeological 
manifestation of the non-Indo-European Etruscans. 

Retreating in time to the next potential intrusion we arrive at the appearance 
of Middle Bronze Age industries in northern Italy, especially the Po Valley. 
Here the emergence of the Terramare culture, the initiation of cremation 
burials into what was previously an inhumation area, the abandonment of 
earlier settlements, and both ceramic and metallurgical parallels with Central 
Europe, particularly Hungary, are all cited as evidence for immigrants. 

Finally, a still earlier series of intrusions has been proposed to explain the 
emergence of the three major Eneolithic/Early Bronze Age cultures of Italy - 
the Remedello, Rinaldone and Gaudo. It has long been argued that these three 
cultures were the result of an invasion(s) of a warrior aristocracy which 
introduced metallurgy, a new burial rite, and new ceramics, as well as a marked 
change in the earlier social system. The archetypal evidence derives from the 
Rinaldone culture of Tuscany which offered evidence of a clear status male 
burial in the Tomb of the Widow at Porte San Pietro. In a stone-cut tomb was 
found a man accompanied by a stone battle-axe, copper daggers, an arrowhead 
and pot, and a woman with evident skull injuries suggesting that she was 
dispatched on the death of her husband according to the rite of suttee as 
practised in ancient India. Other Rinaldone graves offer typically Indo- 
European evidence such as horse remains and an abundance of copper objects 
including daggers, axes, awls and a halberd. To the north, in the Po Valley, was 
the Remedello culture with its large cemetery where metal daggers, halberds, 
axes and awls accompanied burials. The metal is reputed to have come from the 
Alpine foothills of Central Europe; some silver items have also been linked to 
Central Europe. In all of these cultures the physical remains have been 
interpreted as a basically dolichocephalic (long-headed) population among 
whom an intrusive brachycephalic (broad-headed) people settled. 

Venetic: mego zontasto sainatei reitiiai porai egeotora 
Latin; me donavit sanatrici reitiae bonae egetora 

aimoi ke louzerophos. 

(pro) Aemo et (que) liber is. 

'Egetora gave me to the Good Reitia the Healer on behalf of Aemus 
and the children.' 

64 A Venetic inscription from a bronze nail. Here it is translated into Latin and English. 



These arguments for intrusion have, as all others, been attacked and the 
continuity with earlier Neolithic traditions and the lack of clear external sources 
for invasion have been emphasized. A middle ground holds to some small 
intrusion but mainly to continuity among the inhabitants, although this school 
does attribute social change to the introduction and exploitation of metals. 

These different waves of intrusion are still major discussion points of Italian 
archaeology and no theory can be considered as holding the upper hand for any 
period. Obviously there are problems even if one accepts some of the intrusions. 
For example, to suggest that the Rinaldone culture of about 2700 bc 
represented Indo-Europeans derived either from Central Europe or along the 
Mediterranean coasts runs into the obvious problem that it explains Indo- 
Europeans in a territory that emerges later as Etruscan. No one has managed to 
construct a fully integrated model of successive intrusions, from both the north 
and the east, w hich provides a perfect fit with the successive series of different 
Indo-European languages in Italy. While the Adriatic coast combines 
archaeological evidence with a certain degree of linguistic plausibility, our 
search for the earlier Oscans, Umbrians and Latins remains far more elusive. 
That Italy was Indo-Europeanized from some time after 3000 bc and prior to 
800 bc there is little doubt. That one can also postulate a number of intrusions 
from outside Italy and still hold to an essential continuity of cultural 
development in many areas of Italy is also defensible. In a region w here both the 
amount of archaeological evidence and our models for interpreting it still leave 
very much to be desired, this is perhaps all that we can expect at present. 



The Celts 

Although Celtic speakers are now confined to the Atlantic periphery of Europe, 
their Iron Age ancestors once dominated Western and Central Europe, 
occupied vast stretches of Eastern Europe, invaded both Italy and Greece, 
colonized central Anatolia, and, in the guise of mercenaries, even fought in the 
armies of Egypt. Their history during the last five centuries bc is that of a 
rapidly expanding host of different tribes and tribal confederations and it is 
small wonder that the Greeks counted them as one of the great barbarian ethnoi 
of the ancient World, although today they represent the smallest surviving 
group of Indo-European speakers. In examining their origins, we will first 
briefly review the evidence for the different Celtic languages, seek out and 
exclude those areas where we know the Celts to be demonstrably intrusive, and 
then attempt to trace them back to their earliest appearance in the 
archaeological record. 

The Celtic languages are traditionally divided into two major geographical 
groups - Continental and Insular. The Continental languages are those 
recorded in the first centuries bc. Our evidence includes inscriptions in the 
Greek, Latin or Iberian scripts; coin inscriptions; place names and, naturally, 
the personal names preserved in the works of classical historians such as Caesar. 
Most of our evidence for the Continental Celtic languages is assigned to three 
major linguistic groups. There is Gaulish, spoken in the province of the same 
name, and evidenced by nearly 100 inscriptions, the majority of which are 
confined to southern France. There is also Lepontic, a Celtic language known 
from about 70 inscriptions which derives from the Alpine region to the north of 
Milan. The third is Hispano-Celtic (or Celtiberian) whose inscriptions are 
localized to a triangle drawn between Zaragoza, Burgos and Guadalajara, but 
by way of coins and place-names is attested over the northern two-thirds of the 

languages - G aide lie and Brit tonic - are 
indicated with diagonal hatching. 

6 j The distribution of the Celtic 

languages. The continental languages are 
indicated with horizontal hatching, while 
the two groups of insular Celtic 



Iberian peninsula. None of these inscriptions predates the third century bc. A 
fourth linguistic group, the Celtic languages of Eastern Europe, is meagrely 
noted in place and personal names. The Continental Celtic languages became 
extinct largely through the expansion of the Roman empire or the southerly 
movements of Germanic tribes. 

The Insular Celtic languages are found in Great Britain, Ireland and, by way 
of transplantation, Brittany. Although we possess a little very early evidence 
such as the sixth-century BC labelling of Britain as Albion and Ireland as Ierne 
(modern Irish Eire), most of our linguistic data does not begin until the 
incorporation of Britain into the Roman empire. Our earliest major source for 
Ireland is Ptolemy's gazetteer of the known world which dates to about the 
second century ad and provides over fifty names of peoples and places of 
Ireland. By the middle of the first millennium ad the Irish had developed their 
own script (ogham) for inscribing grave memorials. It was also during this time 
that the Gaelic language of Ireland w as carried to both Scotland and the Isle of 
Man. The Brittonic languages of southern Britain survived the Roman 
conquest to evolve into Welsh, Cornish and Breton, the last of which was at least 
partly the product of British refugees who fled to the continent. Today there are 
perhaps three and a half million speakers of the various Insular Celtic 

Our starting point for discussing the expansion of the Celts is their equation 
with the La Tene period of Western Europe which flourished during the last 
five centuries bc. The coincidence of the evidence of historical sources which 
begin by the sixth century BC, the subsequent distribution of Celtic written 

Old Irish 

Froech mace Idaith di Chonnachtaib, mac-side do Be Find a ssidib, 
derbsiur-side do Boind. Is e laech as aildem roboi di feraib Erenn ocus 
Alban, acht nibo suthain. Dobert a mathair di bai dec do assint sid, it e 
finda auderga. Boi trebad occo co cenn ocht mbliadnae cen tabairt 
mna cucai. Coica mace rig rop e lin a theglaig comais cutrummai friss 
uili eter cruth ocus ecosc. 

Froech son of Idath of the Connachta (was) son to Be Find of the fairy- 
mounds, (and) herself sister to Boand. He was the warrior who was the 
most handsome of the men of Ireland and Scotland, but he was not 
long-lived. His mother gave to him 12 cows from the fairy mound, 
white with red ears. He maintained his house for eight years without 
taking a wife to himself. Fifty sons of kings were the number of his 
household, all similar to him in age, in form and in appearance 

68 An excerpt from the Old Irish tale Tain Bo Froich ' Cattle Raul of Froech'. Froech s 
mother 'Be Find 1 is literally 'woman blonde where be is denied from Proto-Indo-European 
*g vv ena- 'woman (cognate with Sanskrit gana, Greek gyne. Old Prussian genna, and English 
queen where it underwent a special semantic development ) . The derb-siur is the 'true-sister', 
the latter element of which is cognate with the Indo-European words for sister, e.g., Sanskrit 
svasar-, Greek eor, Latin soror, English sister. Bound derives from bo 'cow' and find 'white' 
where bo falls within the series of cognate Indo-European words f or the cow: Sanskrit gaus, 
Greek bous, Umhrian bum, Latvian guovs, Tocharian ko and English cow. 


/ f top) The Hit tit e king T udhaliyas IV f 1250-1220 bc) depicted 
111 the embrace of the god Sarumtna on the wall of a rock chamber at the cult site 
of Yazilikaya. The divinity was a Hurrian god adopted by the Hittites. 

2 f above) T welve gods marching in procession from the Hittite cult 
site of Yazilikaya. The physical features of the divinities hooked noses, 
square faces, high cheek bones - have been considered typical of the Hittite 


J Eighth-century jug from a child grave at Gordion, the capital of the Phrygians. 

6 (right) The expansion of the Iranians 
to the west is most dramatically emphasized in this 
relief of a Sarmatian cavalryman carrying a battle 
standard (windsoch). From a funeral stele in 
Chester, one of the Sarmatian veteran camps in 

4 (above) Silver plaque from Luristan. 
The central figure is commonly interpreted as the 
Iranian primeval god Zurvan who produced the two 
eompeting forces of later Zoroastrian religion - 
Ahuramazda and Ahriman. The depletion of young, 
adult and aged groups of people further emphasizes 
Zurvan s character as the god of time. The plaque 
dates to about the eighth or seventh century sc. 

j (left) A Scythian strings a bow on a 
golden vase from the tomb at Kul Oba. The scene is 
one of four and may depict Scythes, the eponymous 
ancestor of the Scythians, in his success/ill stringing 
ofHerakles bow. Note that he carries his own bow in 
typical Scythian fashion in his gory tits, or bow case, 

fastened to his belt. Fourth century BC. 

7 Seal impres- 
sion from \ lohenjo-duro. 
These seals, written in the 
Indus script, provide the 
only information for the 
language of the Indus 
Valley civilization. Still 
undeeiphered, possibly 
undecipherable, the lan- 
guage of the sc ripts most 
likely belongs to the 

8,q Two Vedic gods of the early 
Indo-Aryans. Above is Agni, the god of 
fire and the hearth, whose name is cognate 
with Latin ignis fire', found in English 
'ignite* . On the left is Indi a, the arc hetypal 
god of war, who is commonly depicted 
holding his c lub, the vajra. 

to (tap) A detail of the ' Warrior Vase from Mycenae which 
depicts a troop of bearded Greek warriors wearing horned helmets. The peculiar 
devices hanging from their spears have been variously interpreted as standards, 

slings or even bags for carrying provisions. Twelfth century BC. 

11 (above) Sixteenth-century BC stele from Grave Circle A at 
Mycenae. The scene has been variously interpreted as a battle scene between an 
armed charioteer and a foot soldier, or a chariot race customary in funeral 
games such as those described in the Iliad. In the latter interpretation, (he 

standing figure is an umpire monitoring the turn of the chariot. 

12 Silver-gilt Thracian plaque of the mid-fourth century BCfrom 
Letnitsa. The figure is probably female (Thracian women wore their hair 
shorter than males who wore their hair in a top-knot ) and she may be feeding a 
three-headed serpent, a motif known widely in Indo-European tradition. 

rj (above) lllyrian cavalry 
fighting depu ted on a seventh-century 
BC belt from I ace, Slovenia. Xot 
only are the weapons and clothing of the 
two warriors different but even the 
manes of the horses vary. 


14 (right ) lllyrian warrior 
armed with axe, double spears and 
shield. The figure is from a bronze 
si tula from lace, Yugoslavia and 
dates to the seventh or sixth century BC. 

1$ (above) A silver phi que from Mora- 
via de puis a Slav nobleman on horseback and with 
his falcon. 

16 (right) Face-urn from about the fifth 
century BCfrotn northern Poland. The tradition of 
decorating urns with human faces began earlier in 
Central Europe and was then adopted by the 
prehistoric Baits. 


6g The La Tine culture and the course 
of Celtic expansions in the first 
millennium BC. 

testimony, place names and the Celtic character of the La Tene culture is not 
challenged. Consequently, the expansion of La Tene material culture into areas 
peripheral to its earlier distribution is assigned to the historically attested 
migrations of the Celts. In Hungary, Romania and Yugoslavia, for example, 
there is abundant evidence of intrusive Celtic finds, both metallic and ceramic, 
coupled with Celtic cemeteries and occasionally fortified settlements. These 
began to appear in the fourth century bc when Greek authors record the 
expansion of the Celts towards Southeast Europe. Similarly, Celtic migrations 
and temporary conquests in Italy are seen in the distribution of La Tene 
artefacts in the peninsula. Hence the expansion of the Celts to Eastern Europe 
and the central Mediterranean can all be broadly associated with the 
appearance of La Tene material in these regions. 

Unlike Eastern Europe, the evidence for Celtic expansions along the Atlantic 
pose more problems. We have seen, for example, that Celtic languages were 
established in the Iberian peninsula. The earliest historical attestation for the 
Celts in Iberia is the fifth century bc, for what few earlier sources exist lack 
references to conclusively Celtic populations. They do, however, assist us in the 
central problem of our survey since they help to exclude Iberia as an early 
location of Indo-Europeans. This is evident not only from the inscriptional 
evidence of at least two major non-Indo-European entities in the Iberian 
peninsula - Iberian, attested along the eastern coast of Spain, and Tartessian 
which is known from inscriptions covering the southern quarter of Iberia - and 
also from the only surviving non-Indo-European language of Western Europe, 
Basque. The Basque language appears to have been spoken over a much larger 
area than its present confinement to the Pyrenees would now indicate, and 
place-name studies of northern Spain and southern France exclude this region 
as an early home of the Indo-Europeans. With non-Indo-Europeans well 
attested in the north, east and south of Iberia, the logical consequence is to 
consider this region with those other areas that require an intrusion to explain 
the presence of Indo-European speakers. The absence of substantial La Tene 



influences in Iberia indicates that the intrusion responsible for the Celts must 
have occurred earlier in the period from about iooo to 500 bc when there are 
more abundant associations between Iberia and the Urnfield and Hallstatt 
cultures of France, w hich are regarded as ancestral to the La Tene. There is 
some linguistic support for this hypothesis since the Celtic language spoken in 
Spain appears to have avoided the famous (though perhaps phonologically 
trivial) shift from Proto-Celtic q to p that occurs in all of the other Continental 
Celtic languages. 

The expansion of the Celts to the British Isles poses its own special problems. 
There are no remnant pre-Celtic languages in the British Isles although Pictish, 
the language recorded in inscriptions over southeastern Scotland, may have 
retained a non-Indo-European element in what little of its vocabulary is 
recorded. Otherwise, the earliest evidence for Indo-Europeans in Britain or 
Ireland are the Celts. As the concern of our study is Indo-European origins 
rather than Celtic arrivals, we can avoid the archaeological minutiae involved in 
tracing Celtic migrations into Britain and Ireland and speak in broader terms. 
General archaeological and linguistic opinion assigns the intrusions which 
carried the Celtic languages into Britain and Ireland to sometime during the 
first millennium bc, although some scholars still hold to an earlier date. 
Certainly the similarity between the earliest evidence for Brittonic and Ogham 
Irish are too close to permit them a long separation in time, and they share the 
same Late Bronze Age and Iron Age vocabularies of their continental relations. 

General opinion, therefore, traces the earliest historical Celts back to the 
continent and the La Tene culture, or to its immediate predecessor, the 
Hallstatt culture, at least in Western Europe. 19 In so doing, we can trace the 
earliest Celts to a broad band stretching from eastern France to Bohemia from 
800 bc onwards. Since it is with little difficulty that archaeologists can trace the 
Hallstatt back to the Urnfield culture ( 1 200-800 b c) or yet earlier periods, some 
prehistorians have glibly asserted that a Troto-Celtic' culture can be discerned 
all the way back to the Early Bronze Age. This can only be done, however, if one 
maintains a blissful ignorance that a proto-language, at least to a linguist, must 
meet certain minimal linguistic requirements. In the case of Proto-Celtic we are 

70 Non-Indo-European languages and 
I/ispcino-Celtic in the Iberian peninsula. 



Q and P Celtic 


Old Irish 



k w etwores 








k w er- 



form ' 









7/ The Celtic languages are often divided into Q and P Celtic since Proto-Indo-European *k* 
yielded a T sound (written c in Irish), but a >' sound in Gaulish and Brittonic. 

talking of a state in the linguistic continuum, involving both the phonetic and 
grammatical changes which occurred after the dissolution of Proto-Indo- 
European, but which is immediately ancestral to the known Celtic languages. 
Proto-Celtic is, in a sense, a process whose beginning cannot be precisely dated, 
but linguists would be increasingly uncomfortable with the use of the term as 
one recedes back in time from the end of the second millennium bc. Hence, 
although the archaeologist may perceive continuity in the archaeological 
record, the linguist may prefer to employ less specific terms to describe the 
language of these distant ancestors of the Celts. Since this is the natural 
direction of our own inquiry, we will now leave the trail of the individual Indo- 
European groups and see if there are any broad patterns that may help to 
explain the origins of the Europeans in general. 

Earlier Configurations 

In our review of the origins of the Indo-European languages of Europe, it has 
become apparent that while some of the various languages are intrusive into 
non-Indo-European territories, for example, Greek, Italic and Hispano-Celtic, 
there are no grounds for seeking the immediate origins of any Indo-European 
group outside Europe. We have traced the individual Indo-European 
languages back to an area that includes eastern France and Holland on the west, 
southern Scandinavia and the Baltic Sea on the north, the upper Volga and 
Dnieper on the east, and the territory north of both the Alps and peninsular 
Greece on the south. Throughout the third and much of the second millennium 
BC it is likely that this territory consisted of Indo-European dialects that stood 
somewhere between the language we reconstruct as Proto-Indo-European and 
the 'branches' of Proto-Celtic, Proto-Germanic and so on. It is now our object 
to examine briefly the interrelationships between these different language 
groups and discover if they have any widely accepted expression in the 
archaeological record. 

In general, the dialectal relationships among the various Indo-European 
languages are congruent with their earliest geographical locations. Similarities 
that link two or more languages, the isoglosses of linguistic science, are 
frequent, and provide the type of evidence that suggests that various groups 
were once geographically proximate to one another. When the Dative-Ablative 



ending *-bhyos (Sanskrit vrkebhyas 'to the wolves') appears as *-mos in 
Germanic (Gothic wulfam), Baltic (Lithuanian vilkams) and Slavic (Old 
Church Slavonic vlikomu), this argues for some form of close association 
between the ancestral speakers of these different languages which accounts for 
why they all adopted the same dialectal form. On the basis of this and much 
more abundant evidence of phonetics, grammar and vocabulary, we can speak 
of a continuum of Slavic-Baltic-Germanic that stretched over Northern 
Europe from east to west. We have already seen how Slavic on the east shares 
obvious correspondences or contacts with Iranian, while Germanic displays 
close contacts via loan words with Celtic. It has long been argued that Celtic and 
Italic also share many similarities of phonetics and grammar and even an Italo- 
Celtic branch of Indo-European has been suggested along the lines of Indo- 
Iranian. This is now much disputed, although Italic is regularly included with 
the other West European languages. 

If an archaeologist is set the problem of examining the archaeological record 
for a cultural horizon that is both suitably early and of reasonable uniformity to 
postulate as the common prehistoric ancestor of the later Celtic, Germanic, 
Baltic, Slavic, and possibly some of the Indo-European languages of Italy, then 
the history of research indicates that the candidate will normally be the Corded 
Ware culture. At about 3200-2300 bc this Corded Ware horizon is sufficiently 
early to predate the emergence of any of the specific proto-languages. In 
addition, it is universally accepted as the common component if not the very 
basis of the later Bronze Age cultures that are specifically identified with the 
different proto-languages. Furthermore, its geographical distribution from 
Holland and Switzerland on the west across Northern and Central Europe to 
the upper Volga and middle Dnieper encompasses all of those areas which we 
have seen assigned as the 'homelands' of these European proto-languages. 

Although the Corded Ware horizon may provide a plausible foundation for a 
number of the Indo-European groups in Europe, it does not account for all of 
them. We have also seen how Greek, Illyrian, Thracian, probably Messapic and 
East Italic, and possibly some of the other languages of Italy appear to derive 
from Southeast Europe. In addition, it is from this region that one normally 
obtains the intrusive Phrygians and Armenians who appear in Western Asia. 
The linguistic evidence suggests that Greek, Armenian, Iranian and Indie 
share certain similarities which indicate a chain of languages stretching from 
the Balkans across the Pontic and on into Central Asia. There is some evidence 
that Thracian shares some similarities with the eastern satem languages as well 
as Phrygian. Illyrian and the languages attested from Adriatic Italy tend to 
relate more easily with the western languages. Whatever the precise relation- 
ships, this evidence does not remove these languages from a staging area or 
earlier home in Southeast Europe, an area which lies essentially out of the 
distribution of the Corded Ware horizon. 

Since we have already proposed migrations from the Balkans into Greece as 
early as the third millennium bc, logic compels us to assume that by this time 
Southeast Europe was already a source of Indo-European languages. This is 


72 Most of the Indo-European languages of Europe and Anatolia may be traced back to the 
earlier territories of the Corded Ware horizon (indicated by vertical hatching) and the Balkan- 
Da nubia n complex (dotted line). 

certainly not at variance with majority opinion among Southeast European 
archaeologists who seek the earliest appearance of Indo-European speakers in 
the discontinuity that follows the Late Neolithic/Eneolithic cultures of the 
region. The date for this must be very generally set to about 3500 BC and 
includes a wide variety of local cultures, for example, Maliq III in Albania; 
Karanovo VII or Ezero in Bulgaria; and Baden-Kostolac in the west Balkans. 
The Balkan-Danubian complex has been proposed as a convenient label for all 
of these cultures, and it does provide a basis after which continuity is the 
dominant theme of the archaeological record. 

We may anticipate then that both the Corded Ware horizon and the Balkan- 
Danubian complex are essential to any explanation of the origin of the Indo- 
Europeans of Europe. Both the specific character of these different cultures and 
the problem of their ultimate origins - be they indigenous or intrusive - will be 
the topics of Chapter Eight when we attempt to trace the expansion 
of the Indo-Europeans. Before we can do this we shall have to abandon 
our historical evidence and seek from other sources what the earliest Indo- 
Europeans held in common. 



Proto-Indo-European Culture 

When we have only the reconstructed 
protolanguage, however, we still have a glorious 
artifact, one which is far more precious than 
anything an archaeologist can ever hope to 

Mary Haas, 1969 

Traditionally, there have been two methods employed by Indo-Europeanists to 
reconstruct Proto-Indo-European culture. One involves the straightforward 
comparison of the cultural traits or practices of the different Indo-European 
peoples in the hope that we can isolate common elements and project them back 
to the Proto-Indo-European period. This technique is voluminously illustrated 
throughout the pages of most general handbooks of Indo-European culture 
where the authors amass numerous references to the behaviour and institutions 
of the different Indo-European peoples. Often this makes for the best of 
reading but the entire logic of such an approach, at least when applied to the 
more obviously functional categories of culture, is certainly suspect. 

We may take a familiar example by examining briefly how Indo- 
Europeanists have long observed similarities between the organization and 
behaviour of the war-bands (Manner bunde) depicted in the histories and 
literature of various Indo-European peoples. Here we find, from India to 
Germany and Ireland, a series of recurrent motifs in the organization of these 
warrior sodalities - egalitarian structure, frenzied berserker-like behaviour in 
war and sometimes in peace, the use of wild animals such as wolves as totems, 
and a tendency to operate outside the normal jurisdiction of society which often 
leads to conflicts between the warriors and the formal political and religious 
elites of the community. 

Does such evidence permit us to extrapolate such warrior sodalities back to 
Proto-Indo-European society? Many have certainly found the image of young 
Indo-European berserkers sweeping across both Europe and Asia as attractive 
vehicles for carrying out the expansion of the Indo-European languages. 
Naturally, no one argues that such warrior sodalities are exclusively Indo- 
European and we can cite many other examples drawn from Asia, Africa and the 
New World, especially among the Plains Indians. Nevertheless, if the warrior 
sodalities constituted a formal segment of Proto-Indo-European society, this 
might well be translated into certain archaeological expectations such as the 
systematic deposition of weapons with the burials of young males. Too often, 
this has prompted archaeologists to equate the discovery of warrior burials 


throughout most of Eurasia with the traces of Indo-European expansions. 

The very fact that war-bands are by no means a uniquely Indo-European 
phenomenon should caution us against reading into the archaeological record 
an Indo-European behind every burial accompanied by a stone or metal battle- 
axe. Warfare is the product of environmental, economic and social circum- 
stances that can be found anywhere, and there is no reason for assuming an 
inherently warlike character for the Indo-Europeans. 

More importantly, our evidence for Indo-European war-bands is derived 
from very different time periods with none before the Late Bronze Age. We 
may be struck by the similarities between the heroes of medieval Germanic and 
Irish tales and the Indie war-god Indra who leads his band of hell-raising 
Maruts through the hymns of the Rig Veda, but their behaviour is more apt to 
be generic responses to their particular cultural circumstances than the direct 
genetic inheritance from common ancestors and institutions which existed 
thousands of years earlier. To suggest otherwise is to assume implicitly that the 
structural organization of warfare among the Indo-Europeans remained 
essentially static for several thousands of years. Any archaeologist engaged in 
the study of warfare in Europe could not fail to remark on the numerous 
changes in war technology, defensive architecture and the organization of 
warfare from the Late Neolithic to the Dark Ages, and any attempt to read 
medieval Irish military institutions into the Eneolithic of Western Europe 
would be transparently fallacious. In short, we cannot be entirely confident in 
our reconstructions when they are based solely on the ethnographic residue of 
later Indo-European peoples. Consequently, in this chapter we will keep to the 
second method of cultural reconstruction, linguistic palaeontology. 

Long before August Schleicher had initiated the reconstruction of Proto- 
Indo-European linguistic forms, linguists had already begun to reconstruct the 
culture of its speakers. The same correspondences which demonstrated the 
affinity of the different Indo-European languages also pointed to the shared 
cultural content of the Proto-Indo-European vocabulary. The series of words 
for sheep seen in Luwian hawi-, Sanskrit avis, Greek o(w)is, Latin ovis, 
Lithuanian avis, Old Irish oi or English ewe provided ample proof that the 
Proto-Indo-European community knew the *owis 'sheep'. It was from just such 
comparisons that Adalbert Kuhn, in 1845, attempted to produce a capsule 
description of Proto-Indo-European society. He described the culture of the 
original Indo-European speakers as settled (words for village, fort, house); 
engaged in both agriculture (grain) and stockbreeding (cattle, sheep, goat, pig, 
horse, dog); and politically evolved to the level of the state (king). 

We have advanced far beyond Kuhn's first attempts, and a century and a half 
of lexico-cultural reconstruction has produced a vast amount of research 
including whole encyclopaedias of Indo-European culture. But any survey of 
these sources would soon make it clear that agreement on some issues is difficult 
to find. The reasons for this are sound, if not a bit daunting. 

It is extremely uncommon, for example, for the majority of Indo-European 
languages to share the remnants of the same Proto-Indo-European word. Loss 



of the 'original' vocabulary seems to have been high and is especially likely to 
have affected languages only known in written form within the past 1,000-2,000 
years. If this is the case, in how many different languages must the same word 
occur to be counted as a Proto-Indo-European word? There is really no wholly 
acceptab\e configuration of correspondences that may be utilized, although one 
general rule of thumb demands at least a shared correspondence between a 
European and a non-adjacent Asian language in order to attribute the word to 
high Indo-European antiquity. Others might assign different criteria such as 
correspondences from any three languages provided that at least one was not 
immediately adjacent to the others. For reasons that will become more obvious 
below, it is prudent to demand both a European and an Asian cognate. 

A second problem frequently encountered is the variation in the meaning of a 
cognate word in various languages where it is far easier to reconstruct the 
original sound of the word than what it actually meant. When the Greek word 
for oak is the same as the Germanic word for beech and the Russian word for 
elder, on what grounds does one ascribe an original Proto-Indo-European 
meaning to the word? 

Loan words pose another obvious problem, although by no means so great as 
some have imagined. In the history of Indo-European studies, there can always 
be found a few who challenge the entire validity of our lexico-cultural 
reconstructions since they maintain that it is impossible to know whether a 
particular word has been inherited from Proto-Indo-European into the various 
daughter languages or whether it has merely been borrowed from one language 
to another through time. All too frequently such warnings are capped with 
some example wherein the naive linguist blunders into ascribing coffee, cigars 
and coca-cola to the Proto-Indo-Europeans because of the transparent 
similarity of these words in the modern European languages. 

In a sense, any new word is a loan word which spreads from a single speaker 
or small group of individuals to all other speakers of a language. If these all 
converse in the same idiom, then the word will be accepted and pass for native. 
If, on the other hand, the word crosses a language border, it will be articulated 
according to the rules of the language borrowing the word. When this is 
different from the donor language, then one is often able to discern that it is a 
loan rather than inherited word. In English, for example, we have two words for 
cattle - cow and bovine - that we might wish to relate to similar words in other 
Indo-European languages. We could easily set our word bovine against the 
series that includes Greek bous, Latin bos, and Old Irish bo but we would never 
mistake it for the inherited English outcome of Proto-Indo-European g^ous 
since Proto-Indo-European g® could never produce a b in English (nor would 
the ending of the word be explained). Only cow can be the inherited form, while 
bovine (and beef) are clearly loan words ultimately derived from Latin bos/ 
bovem. Moreover, even bos is not the expected Latin outcome of the Proto- 
Indo-European word (which should have yielded something like *vos in Latin), 
and so linguists regularly regard bos as a loan either from Umbrian where we do 
find the initial b (bum, cf. Latin bovem), or from some similar non-Latin dialect 



of Italy. Even when we are dealing with closely related languages such as Old 
Norse and Old English, we can discern the numerous Norse loan words in 
English, for example, egg, ugly, keel, sky, skill and many others. It ifidiis ability 
to recognize when loan words are present that generally inspires in linguists the 
confidence to determine whether words are inherited from Proto-Indo- 
European or are later borrowings between already differentiated Indo- 
European languages. Indeed, linguists have devoted a substantial amount of 
research into identifying later loan words as a means of elucidating the contacts 
between the different Indo-European languages in prehistory. To take a 
familiar example, we can identify Celtic loan words for iron and lead in 
Germanic which fit neatly with the archaeological evidence that indicates 
contacts between Iron Age Celts and their northern neighbours. Similarly, the 
Germans also borrowed certain social terms, for example, ruler and servant, 
from their Celtic neighbours. This is not to say, of course, that there are not 
genuine difficulties in the analysis of some words, but historical linguists are not 
naive in this and they bring to their data an arsenal of techniques which greatly 
reduces the chance of reconstructing nonsense within prehistory. 20 

Unfortunately, our confidence in the reconstruction of cultural items often 
tends to be inversely proportional to their archaeological utility. For example, 
while no one would doubt that the linguistic evidence indicates that the earliest 
Indo-Europeans knew the dog, it would be vastly more useful if we were certain 
whether they knew the eel, the turtle, the salmon, or other more geographically 
circumscribed species. The reason for our inability to recover with certainty 
some of these words lies embedded in the basic Indo-European hypothesis. An 
expansion of Indo-European speakers over a vast area took many of them out of 
their earlier environment so that they experienced radical changes in their 
cultural ecology before they emerged into history. By this time they had often 
abandoned those parts of their vocabulary that they no longer needed and the 
remaining trace of a particular Proto-Indo-European word may be left in only a 
handful of languages. These may be preserving an old inherited word, but they 
may also be later creations confined to a particular area of the Indo-European- 
speaking world. This is particularly true when we consider the contrast 
between the European and the Indo-Iranian languages. 

If we employ the traditional procedure of not admitting a word as proto- 
Indo-European unless it has at least one European and one Asian reflex, then 
we must encounter the Indo-European fault line. The textual remains of 
Tocharian and the Anatolian languages offer only limited opportunities to 
evaluate the cultural content of the Indo-European languages of Asia. Indie and 
Iranian, on the other hand, provide us with a vast body of evidence but also with 
histories uniquely Asian and apparently essentially pastoral. For this reason we 
often find that words reflected in many European languages do not have a reflex 
in Indo-Iranian. Paul Friedrich, for example, presents us with no less than 
eighteen categories of Indo-European trees, but few of these find any reflex in 
either the Indie or Iranian languages. This has been a problem that has troubled 
linguists for well over a century and which has invariably been resolved in one of 



three ways. Some propose that the Europeans retained the Proto-Indo- 
European vocabulary and the Indo-Iranians moved off from a European 
homeland, losing many terms as they immigrated into their new Asian 
environments; alternatively, it was the Indo-Iranians who retained the 
inherited vocabulary while those peoples who moved west into Europe created 
new words in their new surroundings; finally, some would propose a homeland 
large enough for both branches to encompass a great variety of economies and 
environments. For the present it will be better to hold all of these mutually 
conflicting theories in the backs of our minds and preclude no solution to the 
homeland problem. Rather, let us take a brief survey of the more important 
elements of the reconstructed Proto-Indo-European oulture. 


The physical environment of the Indo-Europeans offers us few clues as to 
where they were situated. They knew both plains and mountains, rivers and 
lakes. The weather vacillated enough to give them words for hot and cold, 
including snow and ice. Only three seasons - winter, spring and summer - are 
reconstructible, with a word for autumn lying beyond our powers to recover. 
This evidence has repeatedly been pressed to demonstrate more about the 
Proto-Indo-European environment and economy than the weight of the 
comparisons can carry. It is unequivocally true that winter is the best-attested 
season and that summer is reasonably well reconstructed. The word for spring 
appears to be a somewhat later linguistic formation and consequently some 
have argued for an environment that sees a rapid transition between summer 
and winter, that is, a more southerly environment. The absence of an autumn 
and therefore of a harvest season, has also been embraced by those who imagine 
the ancient Indo-European peoples to have been pastoralists rather than 
agriculturalists who would surely have maintained a word for this season. Since 
British and American English do not commonly agree on the name for autumn 
(or fall), it would be more than risky to put too much credit on such arguments. 

The word for sea is perhaps one of the more problematic. That a word existed 
(*mori) is most certain. However, it seems originally to have meant swamp, 
marsh land or lake, rather than a larger body of open water. In addition, it is 
found only in European languages and not in Indo-Iranian other than Ossetic - 
an Iranian language contiguous to Europe although originating further to the 
east. Some have argued then that the Indo-European community did not 
originally live near the sea - which sea, of course, was the primary issue since 
such an hypothesis would deny both the Baltic and Black Seas, possibly the 
Caspian or indeed any large body of water. On reflection, it should seem 
obvious that this argument is a logical embarrassment since one would be hard 
pressed to set a people down anywhere in Eurasia without finding some large 
body of water conveniently situated to cry out for a name. 

The botanical evidence for the Proto-Indo-European environment is what 
interests us most here. Armed with quaternary pollen diagrams, we should be 



able to employ this evidence to delimit the area of the Proto-Indo-European 
community and learn something of the arboreal culture of the earliest Indo- 
Europeans. Naturally, any method so elegant and simple in concept is generally 
impossible to execute. 

A survey of the standard Indo-European handbooks reveals that anywhere 
from three to eighteen trees have been reconstructed for the Proto-Indo- 
European landscape. The main problem here lies in the paucity of cognate 
names in any of the Asiatic languages and the abundance of correspondence 
that can be found in the European languages. If we adhere to the rule that there 
must be at least one cognate in Europe and one in Asia, then our conception of 
the Proto-Indo-European forest must, of necessity, be limited. The birch is 
perhaps the best attested, and one may suggest that its peculiar and useful bark 
is the reason for its retention in six language groups. The willow would also 
meet the minimum requirements and its lexical association with words 
indicative of intertwining suggests that one of its uses was in the production of 
osiers for plaiting such things as baskets. The elm could be Proto-Indo- 
European and here we have a tree whose use would probably include foddering 
livestock. Ash, another possible Proto-Indo-European term, is often associated 
with words meaning 'spear shaft' which indicate one of its main uses. If one 
accepts a certain amount of special pleading concerning the root *doru- y which 
normally means tree but sometimes specifically oak, then this tree should also 
be set to the Proto-Indo-European landscape. To omit it is difficult given its 
overall utility as fodder, a source for wooden tools, its use in architecture and its 
association with the religion of many Indo-European peoples. Perhaps 
somewhat less solid are attempts to assign the yew and the pine to Proto-Indo- 
European. The former is also found mixed with terms for archery in a number 
of languages and is sometimes associated with ritual or legalistic functions. 

What can we say from such a collection of trees? Actually, very little, since 
they are all reasonably at home over most areas of the northern hemisphere and 
can be found widely in temperate Europe, the Caucasus, Siberia and Asia 
Minor. Only in areas such as Central Asia and the east Caspian would they 
appear to be less at home, or in areas further to the south. 

Not included in the list of Proto-Indo-European trees is the most celebrated 
arboreal item of the Indo-European homeland problem, the beech. We can 
reconstruct a *bhagos on the basis of words in five different groups. But the 
Germanic (Old Norse bok) and Latin (fagus) forms mean 'beech' while the 
Slavic (Russian buzina) indicates 'elder', and Albanian (bunge) and Doric Greek 
(phagos) mean 'oak'. The importance of the word lies in the fact that the beech 
was traditionally confined west of a line (the famous 'beech line') from 
Konigsberg (the Baltic) to Odessa on the Black Sea. Such a distribution 
effectively excluded a Proto-Indo-European home set on the Russian or Asiatic 
steppe. Naturally, without an Asiatic cognate this argument was refuted by 
those who asserted that it was a late European term, not a Proto-Indo-European 
word. Then, an apparent cognate was discovered in Kurdish (buz), a clearly 
Asiatic language of the Iranian branch, which supported a Proto-Indo- 



European acquaintance with the beech. This was later overturned when the 
Kurdish form was shown not to be cognate with the word for beech, but rather 
with the elm (*wyg-). Even if we admit to a Proto-Indo-European 'beech', late 
holocene pollen diagrams show that the beech extended as far east as the Don 
river and that another species of beech spanned the Caucasus, thus robbing the 
term of much of its geographical utility. 21 

Against all of this is a series of fair to excellent European cognates for the 
juniper, poplar, apple, maple, alder, hazel, nut, linden, hornbeam and cherry. 
This provides us with the classic conundrum: was the homeland in a heavily 
forested zone and did the Asiatic Indo-Europeans abandon their arboreal terms 
as they moved through steppe and desert, or was the homeland set in a very 
thinly wooded area and it is the Europeans who have innovated? Alternatively, 
did the homeland straddle both forest and steppe or desert zones? If we could 
answer those questions, the present book would take a distinctly different 
approach from the one it does. In any event, it is clear from the wild animals 
known to the Proto-Indo-European communities, that the environment must 
have included some tree cover. 

More than a hint of forests comes through our survey of the wild mammals 
known to the Proto-Indo-European speakers. Reasonably solid reconstructions 
provide us with the otter and beaver, the wolf, bear, lynx, elk, red deer, hare, 
hedgehog, mouse and possibly roe deer. Without at least riverine forests we 
could not explain the retention of names such as the beaver and the otter, while 
the bear, red deer and elk also indicate that the setting was certainly not entirely 
confined to the open steppe or desert. 

The names of the Indo-European birds pose a particular problem for the 
linguist since they are often imitative of the actual calls of the birds. One of the 
most obvious examples can be seen in the series: Sanskrit kokila-, Greek 
kokkyx, Latin cuculus, Lithuanian kukuoti, Russian kukusa, Middle Irish cuach, 
and English cuckoo, in which the root *kuku is so clearly onomatopoeic that one 
cannot entirely trust the results of such comparisons. The birds most 
commonly reconstructed to Proto-Indo-European include the eagle and 
possibly some other large bird of prey. In addition, the goose, crane and duck 
are also all attested in the strict sense, and we may note how all of these require a 
habitat in or near water. Less certain are the more clearly onomatopoeic names 
of the hen, owl, hoopoe and crow. 

If we turn now to fish, the quantity of names shrinks abruptly and forms 
great controversy. A Proto-Indo-European *loksos has frequently played a part 
in delimiting the Indo-European homeland, since it has often been taken to 
mean exclusively the salmon (Salmo salar) which would pitch the Indo- 
Europeans into the drainages of the Baltic Sea. But there is fairly sound 
evidence that the word is more likely to have referred to the salmon trout and 
hence is of no great geographical utility. The second term, the eel, is also of 
questionable status (linguistically) and has similarly been employed to exclude 
the Pontic-Caspian region. Even if admitted, one may note that the eel actually 
does frequent the major rivers of the Pontic up to the Sea of Azov. Furthermore, 



a correspondence between Sanskrit saphara and Lithuanian sapalas provides 
the minimum requirement for a Proto-Indo-European fish, possibly belonging 
to the carp family. Two Proto-Indo-European words for fish itself also exist, 
and Eric Hamp has suggested that one of them, *p(e)ik-sk 'the speckled one', 
derives from the denotation of the trout. 

Among reptiles and amphibians, the snake is the best reconstructed but this 
is of almost no value since there is no possibility of specifying what type of 
snake. The turtle has occupied a slightly controversial place. It is not sensu 
stricto Proto-Indo-European but has been employed to exclude certain areas of 
Europe such as Scandinav ia. Even if admitted to Proto-Indo-European it still 
would not delimit an area any smaller than most of Eurasia. We can reconstruct 
words for both bee and honey. The native distribution of the honey-bee covers 
most of Europe and North Asia but excludes the desert and steppe regions to 
the east of the Caspian and Aral Seas. A Proto-Indo-European *wobhsa 'wasp' 
is reconstructed from seven different language groups. 

If we try to draw the environmental evidence together and eschew as much 
prejudice towards particular theories as we can, we arrive at a landscape which 
included some trees and certainly enough to provide forest environments for a 
number of wild mammals. A river-bank or lake-side orientation is discerned 
from some of the animals and birds, although in terms of prehistoric settlement 
this is hardly surprising. That a number of the trees such as birch and willow are 
so closely linked with temperate climates does suggest a region of at least 
seasonally cold temperatures. Beyond this we cannot fairly go other than to 
conclude that to set the Proto-Indo-Europeans exclusively in the open steppe 
(and not forest steppe or river valleys), or in a desert region, would seem to be 
incongruous with the Proto-Indo-European vocabulary. 


That the subsistence economy of Proto-Indo-European society was based on 
stockbreeding with some agriculture is impossible to deny. Indeed, some of the 
most widely and best-attested words in the Indo-European languages are those 
which concern domestic animals, and, of these, words relating to cattle are 
probably among the most prolific. 

Cattle raising is well embedded in the Proto-Indo-European vocabulary, 
with no less than three basic terms - cow, ox and steer - all attested within the 
strict sense. In addition, secondary products such as butter, and possibly some 
form of cheese, can also be attributed to the Proto-Indo-European vocabulary 
although these terms may relate to the products of sheep or goats as well. The 
same may also be said for Proto-Indo-European words for meat, marrow and 
herd. The use of oxen for traction is suggested by the reconstructed terms for 
yoke and plough, and, as we shall later see, wheeled vehicles would most likely 
have been drawn by oxen at the time of the earliest Indo-Europeans. 22 The 
importance of cattle is also reflected in a frozen expression 'to drive cattle', 
generally used in the sense of cattle raiding and found in Celtic, Italic and Indo- 



Iranian. Finally, the cow as a special beast of sacrifice is indicated by the 
Sanskrit-Greek correspondence of the word for the special sacrifice of 'one 
hundred cows'. 

Sheep are also spectacularly well attested in the Indo-European languages 
with cognates in no less than nine major language groups. Their function in the 
economy certainly included wool, as both this word and a word for weaving are 
attested. More problematic is their traditional association with the word for 
wealth where we find a Proto-Indo-European *pek- (to comb) frequently 
associated with *peku- (livestock, possessions). It has long been regarded as 
reasonable that there was an irreversible semantic development that led from a 
word 'to comb' and a noun 'sheep' (the woolly animal) to livestock in general 
and finally to wealth, hence German Vieh 'cattle' and English fee. More 
recently, however, this was challenged by the late Emile Benveniste who argued 
that the semantic development should indeed be reversed and begin with a 
concept of 'movable possessions' which, under the influence of later cultural 
developments, was gradually specified to sheep. Whatever one's position on 
this debate, it would be wise not to read into the linguistic evidence any case for 
the economic primacy of sheep in Indo-European society and to note only their 
clearly evidenced presence and their utility as wool-bearing animals. 

The goat has posed perennial problems for linguists since its lexical 
correspondences tend to be between only a few languages and are not nearly so 
well attested as sheep. Nevertheless, there are sufficient correspondences to 
grant it Proto-Indo-European status and, indeed, it would be difficult to 
imagine a stockbreeding regime in Eurasia that did not employ both sheep and 
goat, although the goat might well have been secondary in importance. It is 
associated with words for hides which provides at least one index of its function, 
and dairy products may also be expected. 

The pig has been a major issue of economic reconstruction since, unlike the 
other livestock, it permits more specific economic and environmental 
conclusions. For those who imagine the Proto-Indo-European economy to 
have been that of pastoral nomads, then the domestic pig should have occupied 
the same position of anathema among the Indo-Europeans as it does among 
Semitic speakers. On the other hand, for those who imagine that the Indo- 
European community had a settled agricultural economy, the pig was one 
indication of stability since it is an animal that is normally excluded from any 
regime of pastoral nomadism. The linguistic evidence for 'pig' is unequivocal as 
a Proto-Indo-European *su- is found widely (English swine, Latin sus, Indie 
su-). The term seemed to refer to the pig generically, and the Indo-Iranian 
terms were specifically applied to the wild pig. It is possible to imagine, 
therefore, that the underlying meaning of the word was a wild pig. This opposes 
the other reconstructible word, *porko-, which is clearly associated with 
domestic pigs but limited in its distribution to the European languages. The 
convenient explanation for this state of affairs is that the Proto-Indo-European 
community was acquainted with the wild pig but only upon the migration of its 
European branches into the settled agricultural regimes of Europe did some of ' 


them acquire a word for the domestic pig. Alternatively, of course, one might 
argue for the disappearance of the *porko- among nomadic Indo-Iranians as 
they migrated from their European homes. Fortunately, neither of these 
arguments is necessary since Benveniste has demonstrated that the automatic 
ascription of *su- to a wild pig is not justified because it is applied exclusively to 
the domestic animal in the European languages. Furthermore, linguists 
discovered reflexes of *porko- in Indo-Iranian. Today, therefore, it would be 
difficult to deny the domestic pig a place in the Proto-Indo-European economy 
and the opposition of *su and *porko is interpreted not as that between a wild 
Proto-Indo-European and a domestic European species but as between the 
name of the adult and young form of the domestic species. 

That the horse was known to the Proto-Indo-European community is 
undeniable as can be seen from the impressive series of correspondences: 
Hieroglyphic Luwian a-su-wa; Mitanni a-as-su-us-sa-an-ni 'horse trainer'; 
Sanskrit asva; Avestan aspa-; Tocharian A yuk\ Tocharian B yakwe\ 
Mycenaean i-qo; Greek hippos; Latin equus; Venetic eku-; Old English eoh; 
Gaulish epo-\ Old Irish ech\ while Lithuanian retains the feminine asva 'mare'. 
As all of these words regularly indicate a domestic horse, linguists are generally 
agreed that the Proto-Indo-European *ekwos 'horse' was also domesticated. It 
is not only widely attested in the Indo-European languages but it is about the 
only animal name to figure prominently in the personal names of the earliest 
Indo-Europeans, for example, the Indie Asva-czkm, Old Persian V\st-aspa y 
Greek Hipp-zrkhos and Phil-ippos, Gaulish Epo-pennus and Old English Eo- 
maer. Moreover, the word is also extended to deities such as the divine twins of 
Indie religion, the Asvins, and the Gaulish goddess Epona. We will later 
examine its role in Indo-European mythology and ritual, but for now we merely 
emphasize the degree to which the domestic horse is embedded in the culture of 
the earliest Indo-Europeans. Problematic is the fact that in many areas of 
Eurasia where the domestic horse first appears there were also wild horses 
present, and there is no relict of a lexical opposition between the wild and 
domestic horse. 

Finally, the word for 'dog' is comfortably and uncontroversially ascribed to 
the Proto-Indo-European vocabulary and, because of the linguistically archaic 
structure of the word, it is attributed to the earliest layer of the vocabulary as 
one might expect from the first domesticated animal. 23 

Although lexical correspondences across all Indo-European languages for 
livestock are easily found, the vocabulary related to agriculture is somewhat 
more limited. One commonly accepted word for an unspecified 'grain' is 
attributable to Proto-Indo-European in the strict sense along with words for 'to 
sow' and for some form of grinding instrument such as a quern stone. The word 
for field *agros is well attested, but whereas in the European languages it 
designates a cultivated area, in Indo-Iranian it refers only to an open plain. 
Words for both plough and sickle, however, can be solidly reconstructed to 
Proto-Indo-European. A much larger common agricultural vocabulary is to be 
found confined to the languages of Europe. Here we find many words 



associated with agriculture - ploughshare, seed, grain, mill, furrow, barley and 
millet. And, of course, the usual opposition between an agricultural west and a 
pastoral, nomadic east has frequently been given as the explanation for this 
dichotomy, with the attendant arguments for lexical loss or innovation 
depending on where one wishes to locate the Indo-European homeland. For 
our purposes it will be best to limit our conclusions to the observation that, on 
the basis of the available lexical residue, at least some agriculture, including the 
plough, was known to the Proto-Indo-European community; we can go no 
further than that until we have seen evidence for the rest of the cultural 


Most of the terms concerning settlement and architecture are so generic that 
they do not offer anything but the vaguest image of a Proto-Indo-European 
settlement. Words for house, post, door and door-post are all reasonably well 
founded. The correspondence between Sanskrit vraja 'hurdle' and Old Irish 
fraig 'wattled wall' has indicated to some that we may imagine walls constructed 
from wattling while associated words for wall, clay and dough suggest the use of 
daub. A common word for hearth is also attested. 

When we go beyond the basic sphere of habitation to larger constructs we 
move into a more interesting but also more debatable region. Cognates derived 
from Proto-Indo-European *u>(e)ik- are found across a number of languages 
with meanings as circumscribed as house (Greek (w)otkos) to village (Latin 
vicus, Gothic weihs) and clan (Indie vis-). On the basis of this one might 
assemble the meanings together and derive them from a word which designated 
a small settlement whose members were related - anything from a body of 
houses belonging to an extended family up to a clan. 24 

Finally, we have a common word for a fortified enclosure seen in such diverse 
languages as Sanskrit pur, Greek polis and Lithuanian pilis. Although the 
meaning of some of these words evolved into that of 'city', their original 
reference to merely a fortified high place is reasonably secure (Greek akropolis 
'high fort'), and we may imagine that some form of fortified settlement or refuge 
existed in the Proto-Indo-European landscape. 


A variety of words for different types of pots and bowls are reconstructible to 
Proto-Indo-European and are seen in such comparisons as Sanskrit caru, Old 
Irish coire and Old Norse hverr, although it must be admitted that the lexical 
reconstructions fall far short of the type of description which might prove 
useful to the archaeologist. Only the correspondence between the Tocharian 
expression tseke . . . peke and Latin jingo . . . pingo, 'I form ... I paint', has been 
pressed to the conclusion that the Indo-Europeans employed painted pottery, a 
conclusion by no means guaranteed by such meagre evidence. 



The vocabulary of metallurgy is poorly and controversially represented. The 
primary utilitarian metal would appear to be that ancestral to the series Sanskrit 
ayas 'metal, iron'; Latin aes 'bronze'; Old Norse eir 'bronze, copper 5 ; and 
English ore. Generally, the fundamental meaning of the word has been taken to 
indicate 'copper', and the meanings 'bronze' and 'iron' have been seen as later 
semantic developments. While the time-scales involved with Proto-Indo- 
European would certainly seem to preclude 'iron', they do not necessarily 
exclude 'bronze' as the original referent. But the absence of any common words 
for tin in the Indo-European languages might support the exclusion of bronze 
from Proto-Indo-European culture unless, of course, the Proto-Indo-Euro- 
pean *ayes was imported bronze and the technique of its manufactures was 
unknown to the earliest Indo-Europeans. A second word for copper, apparently 
related to the Proto-Indo-European root *reudh 'red', is also known, but here 
its similarity to Sumerian urud 'copper' has led to much speculation about 
cultural relations between the Indo-Europeans and the Sumerians or possibly 
an intermediate party. Finally, cognate words for whetstone in Sanskrit, Greek 
and Germanic also suggest the use of metals by the Proto-Indo-European 

There are two possibilities for precious metals, both of which are hampered 
by their derivation from roots so productive and so obvious that one is not 
v certain that they were not independent formations rather than relicts of an 
ancestral word. Gold, for example, is closely linked with words meaning bright 
or yellow, while silver is clearly derived from the word for white. Total certainty 
about these metals is difficult to find, although a strong case can be made for the 
Proto-Indo-European community having been acquainted with silver. 

We have already seen that secondary technologies associated with agriculture 
and stock-breeding were known to the Proto-Indo-Europeans. There is 
certainly sufficient evidence to indicate sewing, spinning and weaving of wool. 
In addition to wool, the Proto-Indo-Europeans knew hemp, and a word for flax 
or linen is shared among most European languages. The linguistic evidence 
tells us little about the appearance of Proto-Indo-European clothes (*wes-) 
other than that the word for belt (*yos-) can be attributed to the earliest Indo- 
Europeans. The grinding of grain, and the yoking of animals for traction, are 
evident in the correspondence of Sanskrit yugam, Greek zygon, Latin jugum, 
English yoke and other cognates. The Indie and Greek correspondence clearly 
indicates a yoked pair of animals, and we may have to do with either the plough 
or, more interestingly from a cultural-historical perspective, wheeled vehicles. 

One of the more noteworthy areas of the Proto-Indo-European cultural 
lexicon is that concerning wagon terminology. It includes several words for 
wheel, as well as one for axle, and another for nave. Beyond these generic terms 
we are helpless in trying to recover a more specific image of the construction or 
appearance of the Proto-Indo-European cart or wagon but, as we shall see later, 
its mere existence proves to be an extremely important cultural marker. 25 

In addition to land transport, there also exists the clear series of cognates: 
Sanskrit naus, Greek nam, Latin navis, Old Irish nau, and so on, which indicate 



the existence of a Proto-Indo-European *nau- 'boat'. Means of propulsion is 
limited to a set of cognates indicating oars. 

With regard to Proto-Indo-European weaponry, the most unequivocal 
reconstructions concern the bow, bowstring and arrow, all of which support the 
existence of archery within the Proto-Indo-European community. A thrusting 
weapon such as a dagger is usually postulated on the basis of the cognates 
Sanskrit asis 'sword' and Latin ensis 'sword'. Here again, we are confronted 
with meanings which appear to be too recent to project back into the Proto- 
Indo-European community without assuming that the name of some thrusting 
weapon later developed into the word for sword in Sanskrit and Latin. 
Alternatively, it has been suggested that the root *nsi- meant only 'black' and 
that we have before us merely parallel developments in Sanskrit and Latin for 
an iron tool or weapon. 

Finally, one might have expected that the Proto-Indo-Europeans would 
have left us with better evidence for their word for axe. As it is we have three 
different terms of varying degrees of controversy. Tomas Gamkrelidze and 
Vyachislav Ivanov cite the comparison of Hittite ates- y Old English adesa and 
possibly Sanskrit -adhiti, all indicating an axe. There is also the notorious 
correspondence between Greek pelekeus and Sanskrit parasu- which are 
traditionally compared with Akkadian pillaq-. The Akkadian word actually 
means 'spindle' and 'spike'. This word has very probably been borrowed into 
Greek and Indo-Iranian from a foreign source and everything about the 
structure of the word suggests that it is not a native Proto-Indo-European 
word. The last possible term is the perplexing series of cognates associated with 
Proto-Indo-European *akmon. It is perplexing because the same word yields 
both the meaning of stone (and hammer) and sky. Hans Reichelt attempted to 
explain this double meaning by postulating an Indo-European belief in the sky 
as a stone vault. More recently, J. Peter Maher has proceeded from the 
underlying root meaning of *ak 'sharp' to suggest that the word originally 
indicated a sharp thing, that is, an axe, which would naturally have been 
fashioned from stone (Maher had in mind the classic 'battle-axes' of the Corded 
Ware culture). The same word also provided the basis for terms referring to sky 
and hammers or missiles, frequently associated with Indo-European thunder 
gods. This complex of seemingly ynrelated meanings is resolved along the same 
lines known to anyone familiar with the history of early antiquarian studies. As 
late as the eighteenth century there were still scholars who persisted in the 
common folk-belief that stone axes were the residue of thunderbolts (hence 
thunderbolts, elf-shot in English or Donnerkeil in German). This, Maher 
maintained, was merely the retention of earlier Indo-European beliefs that 
metaphorically associated axes with phenomena such as thunder and lightning. 

Social Organization 

There is a sort of horrible irony in the fact that, while modern archaeologists are 
greatly interested in reconstructing the social systems of prehistoric peoples, 



historical linguists offer the archaeologists such detailed reconstructions that 
they are still beyond archaeological retrieval even when we know what to look 
for. Nothing can make years of attempts to recover social systems on the basis of 
detailed studies of mortuary evidence appear so futile as to be informed by a 
linguist that the Proto-Indo-European community appears to have employed 
an Omaha-type kinship system, since no one knows what this should look like 
'on the ground'. All linguistic evidence suggests that Proto-Indo-European 
society was patrilineal in descent and male dominated according to that much- 
overworked term patriarchal. We lack a common term for husband or wife 
although we can recover a Proto-Indo-European *widhewa 'widow'. We cannot 
reconstruct a common word for marriage for Proto-Indo-European; but as we 
have seen, many Indo-European languages do employ the same Proto-Indo- 
European verb *wedh- 'to lead (home)' when expressing the act of becoming 
married, from the groom's point of view. This suggests that the residence rules 
of the Proto-Indo-Europeans involved the woman going to live in the house of 
her husband or with his family. 

* aw(y)os 
(Fa Fa) 

(Fa Mo) 

* aw(y)os 
(Mo Fa) 

(Mo Mo) 

* patrwyos * pater 

(Fa Bro) (Fa) 

* mater 

* aw(y)os 
(Mo Bro) 


* bh rater 

* sunus 
(Bro So) 

' swesor 

' sunus 

k dhugheter 

(Si So) 

(Si Da) 

* nepots 
(So So) 

* neptis 
(So Da) 

(Da So) 

* neptis 
(Da Da) 

KEY: Fa = father Mo= mother | 
Bro= brother Si = sister 
So = son Da - daughter j 


73 Proto-Indo-European kinship terms reconstructed according to the Omaha system. One of 
the characteristics of the Omaha system is the skewing of generations where the same term is used 
for both grandfather and mother s brother, e.g., *aw(y)os: Hittite huhhas 'Fa Fa, Mo Fa, 
Armenian hav 'FaFa, MoFa, Old Church Slavonic uj 'MoBro', Old Prussian awis 'MoBro\ 
Latin auus 'FaFa, MoFa\ Old Icelandic afi 'FaFa, MoFa\ Old Irish aue 'grSo'; reciprocally, 
the same term is proposed for grandchild and nephew, e.g., *nepots: Sanskrit napat 'grSo, 
descendant*, Avestan napa 'grSo, descendant', Greek nepous 'descendant', Albanian nip 'grSo, 
nephew', Old Church Slavonic netii 'SiSo\ Lithuanian nepuotis 'grSo, SiSo', Old English 
nefa 'grSo, SiSo', Latin nepos 'grSo, SiSo, SiDaSo', Old Irish nie 'SiSo\ and Cornish nie 
'grSo, nephew'. The reconstruction of the underlying Proto-Indo-European meanings are by no 
means universally agreed and many linguists argue that some meanings such as 'nephew for 
*nepots were actually very late developments in the individual Indo-European languages and 
not inherited from the proto-language. 



Among the kinship relations, the roles of the mother's brother and the 
corresponding sister's son is of especial interest. The basic linguistic forms 
appear to be *aw(y)os for both the mother's father and mother's brother and 
*nepots for both grandson and sister's son, a pattern that some argue is 
congruent with the skewing of generations found in the Omaha system. 26 
Throughout the Indo-European world we observe a well-known pattern of 
behaviour predictable from the kinship structure of the Indo-European 
languages. In general, the father occupies the role of stern disciplinarian as do 
father's brothers who, in a patrilineal system, are all competing for positions of 
authority over their younger kinsmen (and potential competitors). On the other 
hand, in a patrilineal system the maternal uncle is outside the lineage and his 
behaviour is generally that of both an affectionate and friendly counsellor. It is 
in this light that the statement bf Tacitus on the early Germans, 'A sister's sons 
are considered to be related to her brothers as nearly as to their own father', and 
other historical references, are to be understood. It follows naturally, then, that 
as a man would have an especially close relationship with his own maternal 
uncle, similarly he would occupy the same role towards his sister's sons, the 
Proto-Indo-European *nepots. Beyond this particular pattern of behaviour, it 
has also been suggested that fosterage outside one's own family appears to be 
strongly associated with this same pattern: for example, Beowulf is raised by his 
maternal grandfather; the Irish champion Cu Chulainn is fostered by his 
maternal uncle Conchobor; Roland and Charlemagne, and so on. As we will see 
in the next chapter, this particular relationship of *nepots to uncle received 
special elaboration in Indo-European mythology. 

Beyond the family, several institutions appear. One is the clan, and its leader 
(*n>eik-potis) is seen in the striking comparison between Avestan vispaitis 'clan- 
chief, household chief, and Lithuanian viespats iord', formerly 'clan-chief, 
together with its very nontransparent Albanian cognate zot iord'. But beyond 
the level of the clan it is more difficult to proceed confidently. We do know that 
in ancient Iran the tribe was designated the zantu, a word with impeccable 
Indo-European origins and coming from the same root which yields words like 
Latin gens 'race, tribe', Old Norse kind 'follower', and English kin but here we 
speak of identity of root and not of word and it is impossible to know whether we 
are dealing with an inherited word for tribe or not. 

Two terms seem to be associated with some form of military organization. 
The most familiar derives from Proto-Indo-European *teuta- which is 
reflected in the Persian word for mob (toda) but is translated more generally as 
'people' in Oscan touto, Old Irish tuath, Latvian tauta, Old High German diota 
(whence Deutsch), arguably Hittite tuzzi-, and in personal names recorded 
among the Greeks, Illyrians and Gauls. Furthermore, there is the correspon- 
dence of Hittite lahha- 'campaign', Mycenaean ra-wa-ke-ta (lawagetas) 
'commander', Greek la(w)os 'people under arms' and Phrygian lawagtaei 
'commander'. To what extent either of these terms can be extrapolated back 
into Proto-Indo-European society to indicate war-bands or more abstract 
conceptualizations of the people under arms is debatable. Nevertheless, few 



would deny that this evidence indicates the existence of some form of military 
institution among the earliest Indo-Europeans. The vocabulary of strife is also 
seen in a Proto-Indo-European word for blood revenge or blood payment 
which is attested in Avestan and Greek. 

The apex of Proto-Indo-European society, according to the standard 
handbooks of Indo-European culture, was ruled by a king whose title has 
usually been secured by the textbook series: Sanskrit raj, Latin rex, Gaulis rix, 
Old Irish n, and possibly Thracian Rhesos. It has been assumed that this Proto- 
Indo-European *reg- was not necessarily the secular ruler whom one might 
normally envisage. Linguists have argued that the root of the noun is *reg which 
provides such meanings as stretch, draw out in a straight line, and straighten. 
Our English word right is a reflex of this root, and the same opposition which we 
employ between what is straight or right and what is bent or crooked, that is, 
dishonest or wrong, is encountered throughout the Indo-European languages 
(see Chapter Five). Jan Gonda and Emile Benveniste sought in the basic 
etymology of the word a hint of the original function of the Proto-Indo- 
European *reg-. Gonda suggested that the word meant one who stretches or 
reaches out, a metaphor for the formal activities of a king who is often depicted 
in Indo-European tradition as fulfilling his duties with outstretched arms. 
Benveniste argued that the fundamental meaning was 'one who determined 
what was right'. This suggested a leader who was more concerned with 
maintaining social and moral order than a secular sovereign exerting coercive 
power over his subjects or leading them in battle. Indeed, Benveniste proposed 
that there may have been more overtly priestly functions associated with the 
Proto-Indo-European king in that the root meaning of 'stretching out' or 
'straightening' might be associated with duties such as laying out limits, be they 
demarcations of sacred territory within a settlement, the settlement itself or the 
borders of national territories. 

This whole concept of Proto-Indo-European kings has recently been 
challenged. First, Andrew Sihler has argued that the underlying root is not 'to 
arrange in a straight line' but 'be efficacious, have mana\ Then Hartmut 
Scharfe reviewed the Vedic evidence, our only body of material providing an 
Asiatic cognate, to discover that the word raj in the earliest Vedic texts was not 
the masculine noun meaning king but a feminine noun indicating 'strength, 
power'. If this is accepted, then we no longer have evidence for Proto-Indo- 
European kings, and our testimony is limited to Celtic and Italic, two languages 
which share numerous similarities and which suggest a particular political 
development among some late Indo-European groups of Western Europe. 
Scharfe does observe that the correspondence between Sanskrit rajan- and 
Greek aregon suggests a Proto-Indo-European word for 'protector' or 'person 
with power or charisma', but not 'king'. The highest socio-political level that 
Scharfe attributes to the Proto-Indo-Europeans is our *weik-potis 'the master 
of the clan' which we have already reviewed above. 

We must also take a brief glance at that most loaded of Indo-European words 
- Aryan. As an ethnic designation, the word is most properly limited to the 



Indo-Iranians, and most justly to the latter where it still gives its name to the 
country Iran (from the Avestan genitive plural aWyanam through later Iranian 
eran to it an). The great Persian king Darius described himself as Aryan. The 
term was also used widely in India where it referred to one who was a member of 
the community (though details of who was included in the community have 
been the topic of wide and unsettled debate). Whether this ethnic designation 
was limited to the Indo-Iranians or not is difficult to say. A possible cognate 
occurs in Hittite, for example, where it indicates 'kinsman, friend', and there 
also appears here the negative expression natta ara 'not proper to the 
community', that is, 'not done'. Although some claim that this root can be 
found in the names of many other Indo-European peoples, for example, Irish 
Eriu and aire, this would require more argument than is worth the effort and we 
are safer to remain with the general consensus that it does not rather than to 
pursue this matter further. 27 


Our review of Proto-Indo-European culture omits volumes that have been 
written about the reconstructed vocabulary, since much falls either under the 
category of predictable phenomena or else under items not readily retrievable by 
the prehistorian from any source other than language. Day, night, earth, sky, 
clouds, sun, moon and star can all be found in the reconstructed Proto-Indo- 
European vocabulary. We may be confident that the Proto-Indo-Europeans 
were physically similar to us and that many of their anatomical parts are 
linguistically retrievable through the comparison of the Indo-European 
languages. Indeed, it is bizarre recompense to the scholar struggling to 
determine whether the Proto-Indo-Europeans were acquainted with some 
extremely diagnostic item of material culture only to find that they were far more 
obliging in passing on to us no less than two words for 'breaking wind'. English 
dictionaries may occasionally shrink from including such vulgar terms as 'fart' 
but the word gains status when set within the series: Sanskrit pardate, Greek 
per do, Lithuanian perdzu, Russian perdet\ Albanian pjerdh 'to fart loudly' 
(distinguished from Proto-Indo-European *pezd- 'to break wind softly'). 

Turning to more crucial matters, we can see that the presence of words for 
pottery, domestic animals and agriculture in the Proto-Indo-European lexicon 
argues that the community was at least Neolithic and that it would be senseless 
to assign the Proto-Indo-Europeans to the earlier hunter-gatherer societies of 
the Mesolithic. Moreover, we encounter in the Proto-Indo-European vocabu- 
lary not only words that one associates with the original 'Neolithic Revolution' 
but also the lexical residue of what Andrew Sherratt has termed a later 
'Secondary Products Revolution'. These include the secondary uses and 
products of domestic animals which can be seen, in dairy products, wool and 
textiles, wheeled vehicles, yokes, ploughs and the domestic horse. Sherratt 
suggests that these secondary products only emerged in Europe several 
thousand years after the initial appearance of the Neolithic economy. 


When we consider the most recent terms of the inherited Indo-European 
vocabulary, most are not evidenced in Eurasia prior to the fourth millennium 
bc. The earliest wheeled vehicles are clearly a fourth-millennium phenomenon 
whether they be initially found in Mesopotamia, the Caucasus, the Pontic- 
Caspian steppe or along the Danube. The horse appears to have been first 
domesticated about the beginning of the fourth millennium and such metals as 
silver are rarely found anywhere in Eurasia prior to the fourth millennium bc. 
Since caution teaches us that the evidence gained from future excavations will 
most probably increase the antiquity of any invention, we might then assign a 
notional date of about 4500 bc as the earliest probable time for the culture 
reconstructed from the inherited vocabulary of the Indo-European languages. 
We have already seen from our review of the evidence for the earliest Indo- 
Europeans in the historical record that a terminal date for the Proto-Indo- 
Europeans would be set not much later than 2500 bc, possibly somewhat 
earlier. In the broadest terms then, the Proto-Indo-Europeans were a Late 
Neolithic or Eneolithic society which began to diverge about 4500 to 2500 bc. 28 
This is the time-frame we will employ when we embark on our quest to discover 
the homeland of the Indo-Europeans. But first we must review yet one more 
source of cultural information that might offer an insight into the nature of 
Proto-Indo-European society. We need to examine briefly the massive 
evidence for the religion and mythology of the Indo-Europeans. 



Indo-European Religion 

The oldest religious rites of Indo-European 
peoples do not presuppose temples or idols. Nor 
is there a reconstructible term for 'temple'. But 
there is a 'worship', conceived as a hospitable 
reception with a meal, consisting of slaughtered 
animals, and accompanying recitation of poetry, 
the 'celestials' coming, as it were, on a v isit to 
the 'earthly ones'. 

In what did the Proto-Indo-Europeans believe, or, to use their own words, to 
what did they 'put their hearts'? This archaic expression is still preserved in a 
roundabout way in English where the Latin verb credo 'I believe' has been 
borrowed to fashion our English creed. This word finds cognates in Old Irish 
cretim, the Hittite karatan dai, Indie srad-dha and Avestan zrazda-. Admittedly 
with some linguistic difficulty, the Proto-Indo-European expression appears to 
have been built from the words for heart (*kerd-) and put (*dhe). In order to 
examine the ideology of the Proto-Indo-Europeans we will require access not 
only to the direct evidence of the reconstructed vocabulary but also to the less- 
tangible evidence of the myths of the various Indo-European peoples. As Indo- 
European mythology has attracted an enormous volume of scholarly interest, 
we will prune the subject down by concentrating our attention primarily where 
the evidence for Indo-European religion may tell us something of the ritual 
behaviour and structure of Proto-Indo-European society. 

Although philologists have long been interested in the religion of the ancient 
Indo-Europeans, the results produced by the type of linguistic reconstruction 
we employed in the last chapter are neither numerous nor always particularly 
solid. One of the more obvious correspondences can be seen in the similarity of 
Sanskrit devas, Latin deus, Lithuanian dievas. Old Irish dia and the Old Norse 
plural tivar 'gods'. In addition, there is that most striking of all comparisons: 

Pall Thieme, 1964 











Illy rian 


D Sius 






Although we can produce a fairly facile translation of Proto-Indo-European 
*dyeus pdter as 'Father Sky', we cannot be confident that we understand the role 
of this god in the religion of the Proto-Indo-European community. Some 
would doubt that 'father' here connoted a progenitor of gods or man but rather, 
as has long been argued, that it signalled only the type of authority which one 
associates with the Latin paterfamilias. Secondly, we find this same god at the 
apex of some Indo-European religions such as Greek and Roman, but of less 
obvious importance in others such as Indie. Some have regarded the 
ascendancy of 'Father Sky' in the Mediterranean to have been a relatively 
recent phenomenon involving the conflation of an earlier Proto-Indo- 
European god with local Mediterranean weather/storm deities. 

Most other lexical correspondences tend to be associated with predictable 
natural phenomena. For example, a sun-god (or goddess) is normally 
postulated on the correspondence of Sanskrit Surya, Gaulish Suits, Lithuanian 
Saule, Germanic Sol and the Slavic Tsar Solnitse. Furthermore, we may 
reconstruct common names for moon and dawn, both of which appear as divine 
figures in various Indo-European religions. A somewhat more difficult 
correspondence may conceal a thunder- or rain-god behind the debatable 
comparison of the Indie rain-god Parjanyas, Lithuanian Perkunas, Slavic 
PerutC and the Norse Fjorgyn, the mother of Thor whose credentials as a 
thunder-god hardly need defending. 

With such correspondences it is small wonder that many Indo-Europeanists 
were content to view the Proto-Indo-European pantheon as little more than the 
theomorphization of the major elements of nature. To press beyond this level of 
comparison required a certain amount of linguistic legerdemain which yielded 
rarely accepted equations. Some, for example, could point to the possible 
linguistic similarity between Kerberos, the guardian dog of the Greek Hades, 
and the epithet sabala 'spotted, varicoloured' (*kerberol), the standard epithet 
of one of the dogs of Yama, the Indie god of the dead. And evenafter more force 
than the comparative method in linguistics will normally allow, all one gains by 
postulating such a correspondence is the somewhat incongruous image of a 
Proto-Indo-European canine guard of the realm of the dead who answered to 
the name of 'Spot!' 

More promising perhaps are those mythological reconstructions whose 
linguistic credentials are reasonable although they do not necessarily provide us 
with a clear image of their place in Proto-Indo-European ideology. Certainly 
one of the more intriguing examples is the comparison of the Indie (and 
Avestan) Apam Napat 'grandson/nephew of water' with Latin Neptunus and 
the Irish Nechtain. The latter two preserve only the element *nepots 'grandson 
or nephew' but were still closely associated with water, the Latin god as the 
Roman equivalent of Poseidon, the god of waters, and the Irish figure Nechtain 
who maintained a sacred well. The *nepots also figures prominently in 
eschatological literature. The epic traditions of a number of Indo-European 
peoples preserve an account of the 'final battle', for example, Kurukshetra in 
the great Indian epic, the Mahabharata\ the 'Second Battle of Mag Tured' 



among the early Irish; Ragnarok among the Norse; and several others. A 
common structure has been found to underlie these different accounts which 
casts the *nepots in the role of the protagonist against his evil opponent. 

We also have the correspondence between the Indie Manu, the ancestor of 
the human race, and the Germanic founder-figure Mannus. The Indie god 
Aryaman, a deification of the concept of Aryan-hood, may share a Celtic 
cognate with the Gaulish Ario-manus and Old Irish Airem. Finally, we may 
note the similarity between the Indie god Bhaga 'sharer or dispenser of goods' 
(which, via its Iranian cognate was lent into the Slavic languages as the word for 
god), and the Phrygian Zeus Bagaios. 

But all of these comparisons have never been regarded as entirely satisfying 
and it seems that straight lexical reconstructions of Proto-Indo-European 
divinities have proved far less rewarding here than they have for the 
reconstruction of other aspects of culture. If any further progress is to be made, 
some argue, then it requires the replacement of a method based on comparative 
philology with a 'new comparative mythology'. 

Dumezil and tripartition 

The foundation of much of what is currently written about Indo-European 
mythology has its origins in the sociological approach to the study of religion 
championed by Emile Durkheim. He and his followers proceeded from the 
assumption that myths expressed certain social and cultural realities. The 
attractiveness of such an approach is immediately apparent to anyone who has 
pondered the range of social structures reflected in the mythologies of various 
peoples. The Sumerians, for example, appear to have venerated a pantheon of 
gods organized according to an archaic version of their own social order. Among 
the Indo-European peoples we find the old Germanic social system and values 
encapsulated in the Old Norse pantheon housed at Valhalla, while the ancient 
Greek Olympus more closely reflected the more complex and specialized 
institutions of early Greek society. Moreover, the myths of a people were not 
only to some extent ciphers of their (often archaic) social structures, but they 
also reinforced social behaviour and served as divine charters for political 
realities. The Scythian origin myth, for example, records how their neighbours, 
the Agathyrsi and Geloni, had to resign themselves to subservience to Scythian 
power since their eponymous ancestors ( Agathyrus and Gelonus) had failed in a 
mythic contest of strength to string Herakles' bow. Scythes, the progenitor of 
the Scythians, had naturally accomplished the task and secured a divine charter 
guaranteeing the superior social position of the Scythians. With such an 
approach, the study of Indo-European religion invites our closest attention 
since the antiquated social realities which might be preserved in the myths of 
different Indo-European peoples may shed some light on the nature of Proto- 
Indo-European society itself. It is obvious, however, that if we are to pursue 
this line of inquiry, we must abandon our hesitation stated at the beginning of 
the last chapter at employing such evidence. Reconstruction of Proto-Indo- 


European ideology by comparing the structures of the myths of different 
peoples without further linguistic support differs little from other cross- 
cultural comparisons similarly obtained. Yet to dismiss such evidence here 
would exclude any discussion of the work of the majority of scholars now 
engaged in the study of Indo-European myth. So much for my procedural 

An obvious starting point to such an investigation is how the various Indo- 
European peoples perceived the social divisions of their own communities. 
Among the earliest attested is the familiar division of society in Vedic India into 
the brahmanas 'priests', ksatriyas 'warriors' and vaisyas 'herder-cultivators', 
with the sudras, the lowest group, outside the Aryan community and composed 
of the suppressed indigenous population. Such a scheme has not only been 
remarkably persistent in India until the present but is quite analogous to the 
social divisions imputed to other Indo-European societies as can be seen in the 
following table: 

India Iran Greece Rome Gaul 

(Ionian bioi) 

I priests brahmanas athravan- priests and flamines druides 


II warriors ksatriyas rathaestar- warriors milites equites 

III herder- vaisyas vastriyo labourers quirites plebes 
cultivators fsuyant- and artisans 

All of this evidence suggests a conceptual framework among early Indo- 
European-speaking peoples that tripartizes society into three classes: priests, 
warriors and herder-cultivators. Is a residue of such a system also recoverable 
from Indo-European mythology? 

An emphatically affirmative answer to this question is given by the eminent 
French comparativist Georges Dumezil and his colleagues and followers. They 
have produce a vast corpus of evidence that has apparently formed a foundation 
for interpretation solid enough to withstand not only its bitterest critics but 
even the frequent excesses of its over-zealous supporters. 

Dumezil argues that the evidence for tripartition of Indo-European society 
can be seen in one of the earliest sources of Indo-European religion - the treaty 
between Matiwaza, King of Mitanni, and the Hittite king, dating to about 1380 
BC and discovered in the archives of Bogazkoy (Hattusa). The Mitanni king, as 
we have seen before, evoked the names of the transparently Indie gods Mitra, 
Varuna, Indra and the Nasatyas. The first two names are characteristically 
found co-joined in the Vedas, that is, Mitra-Varuna, and they represent, 
according to Dumezil, the two main aspects of Indie sovereignty. Mitra 
personifies the concept of contact and governs the legalistic aspect of 
sovereignty while Varuna's domain pertains more appropriately to the magical 
or religious. The god Indra is the warrior-god par excellence while the Nasatyas 
are twins, associated closely with horses, and find their clearest roles in the 
maintenance of health in both livestock and people. In short, the three 


fundamental estates of Indo-European society are presented in canonical order 
in the Mitanni treaty. 

This same tripartite division is seen over and over again throughout the 
mythologies of the Indo-European peoples. Herodotus records, for example, 
how the kingship of the Scythians was awarded to one of three brothers who 
could pick up three heavenly (but burning) objects that fell to the earth - a cup, 
an axe, and a plough with yoke. The first is regarded as a symbol of the ritual 
and sovereign function, the axe is the instrument of war and the plough with 
yoke are clearly symbols of the cultivator. The pre-Capitoline divine trio in 
ancient Rome consisted of the sovereign Jupiter, the war-god Mars and, finally, 
Quirinus, the patron of the people. Or, to take a more well-known example, 
preparatory to the disastrous judgment of Paris in Greek mythology, the three 
goddesses in competition each attempted to bribe Paris with a primary aspect of 
their own character. Hera offered sovereignty, Athena promised military 
prowess while Aphrodite promised the love of the most beautiful woman in the 
world, an arguably obvious aspect of fertility. 

Tripartition is by no means limited to divine figures but permeates other 
aspects of society. Medicine, for example, is also divided when we find that, 
according to Pindar, the Greek healer Asklepios heals sores with spells, wounds 
with incisions and exhaustion with herbs and potions. A similar system is 
encountered in the Iranian Avesta where three types of medicine are listed: 
spell-medicine, knife-medicine and herb-medicine. Diseases and cures per- 
taining to the sovereign class are healed with spells appropriate to the 
techniques of priests. Wounds inflicted in battle, or fractures, are the 
province of surgery. Wasting diseases that threaten general well-being are 
treated with herbs and potions. 

These comparisons are almost limitless and new articles invariably add to the 
number of canonical recitations of Indo-European tripartition. The underlying 
system, according to Dumezil, is one where society is encapsulated in three 
basic elements or, to use the Dumezilian expression, 'functions': 

1 The first function embraces sovereignty and is marked by a priestly stratum of society which 
maintains both magico-religious and legal order. The gods assigned the sovereign function are 
often presented as a pair, each of which reflects a specific aspect: religious such as the Indie 
Varuna or Norse Odinn, and legal such as Mitra or Tyr. 

2 A second military function assigned to the warrior stratum and concerned with the execution 
of both aggressive and defensive force, for example, the war-gods Indra, Mars and Thor. 

3 A third estate conceptualizing fertility or sustenance and embracing the herder-cultivators. 
Here the mythic personages normally take the form of divine twins, intimately associated with 
horses, and accompanied by a female figure, for example, the Indie Asvins (horsemen) and 
Sarasvati, the Greek Castor and Pollux with Helen, the Norse Frey, Freyr and Njorth. 

Although the tripartite conceptual system proposed for the ancient Indo- 
Europeans offers some opportunity for archaeological confirmation, it is a bit 
surprising to see how little use of archaeology has been made by those interested 
in comparative mythology. One of the few exceptions has been an attempt by 



Dumezil himself to analyze one of the Luristan bronzes according to Indo- 
European mythology. The bronze capitol, dated to about the seventh or eighth 
century bc, is illustrated with seven registers, the upper and bottom two of 
which can be dismissed as primarily ornamental. It is the three central friezes 
that offer, according to Dumezil, iconographic evidence of the Indo-European 
system of tripartition. The upper register portrays two figures symbolically co- 
joined by both holding the same palm in the centre. The left hand figure stands 
next to an altar, a clear association with religious functions, while the right hand 
figure stands next to a bovine. Dumezil reminds us that the bull was the titular 
animal of Mitra, and he identifies the two figures as the sovereign gods Varuna 
and Mitra. The middle register depicts a figure standing between two lions with 
a bird overhead. Dumezil suggests that the figure is quite probably of Indra, the 
Indie warrior-god. Of the 36 mentions of a bird in the Rig Veda, 23 of them are 
associated with Indra while another 6 occur with the Maruts, Indra's warrior 
band. Twelve of the 13 mentions of a lion in the Rig Veda are connected with 
Indra or the Maruts. The lower register depicts two figures, interpreted as the 
Indie divine twins, the Asvins, assisting an older figure, an iconographic 
representation of an incident in the Rig Veda where the twins rejuvenate an old 

Whether one is impressed by this interpretation or not, it must be admitted 
that there is ample room for archaeological 'testing' of the tripartite model. 

One of the more obvious symbols of social tripartition is colour, emphasized 
by the fact that both ancient India and Iran expressed the concept of caste with 
the word for colour (varna). A survey of the social significance of different 
colours is fairly clear cut, at least for the first two functions. Indo-Iranian, 
Hittite, Celtic and Latin ritual all assign white to priests and red to the warrior. 
The third function would appear to have been marked by a darker colour such 
as black or blue. Unfortunately, the preservation of coloured textiles among a 
prehistoric people is possible but seldom encountered and one must seek more 
enduring markers of Indo-European social classes. 

Perhaps a potentially more rewarding area for examination can be found 
among the ritual animal sacrifices that we encounter among the early Indo- 
Europeans. The evidence of these rituals, especially those preserved in ancient 
India and Rome, demonstrates how a hierarchy of different victims were 
sacrificed to, or associated with, the various divinities who filled out the major 
social 'functions' of Indo-European mythology. In the Indie sautramani, for 
example, the priestly Sarasvati received a ram, the warlike Indra obtained a 
bull, and the Asvins, the twins who represented the third estate, were offered a 
he-goat. In the Avesta, the great goddes Arad vi Sura Anahita, who embraced all 
three functions, received the sacrifice of horses, cattle and sheep. The Roman 
purification sacrifice of the suovetaurilia preserved within its very name the 
identity of its three ritual victims - su 'pig', ovis 'sheep' and taurus 'bull'. 

Although the sacrificial sequence in these and other rituals was clearly 
hierarchical, the precise identity or sequence of the victims sacrificed was not 
rigidly observed within the same culture, much less between different Indo- 



74-77 Lurisian bronze covering for a 
quiver dating to about the eighth or 
seventh century bc (L. 8.25 cm). 
Georges Dumezil has interpreted the 
figures as representa tives of the three 
Indo-European functions'. Three 
Registers are shown in detail. From top 
to bottom, Register j; The 'sovereign 
figures; Register 4: The 'warrior figure; 
and Register The 'twins . 



European peoples. In examining the Indie evidence, for example, Jaan Puhvel 
notes that where the horse is identified as one of the victims it is dedicated to the 
warrior-god, while a sheep or hornless ram is offered to the priestly deity and 
cattle or goat to those representing the third estate. But where the horse is 
absent from the ritual, cattle replaces it and the third function receives a goat or 
pig. The Roman evidence shows even greater variability, and in the Greek 
triple sacrifice known as the trittua, we find that the animals are often a ram, 
bull and a boar. 

The difficulties involved in extrapolating from this type of evidence to the 
Proto-Indo-Europeans are fairly obvious. While the horse may normally be 
associated with a warrior-deity and the sow is certainly an archetypal third- 
function fertility symbol, we can see how exceedingly difficult it is to assign 
specific socio-ritual identities to the other victims. As Jaan Puhvel observes, the 
important factor may not necessarily rest in any inherent symbolism associated 
with a particular species but rather to whom they are to be sacrificed, different 
gods requiring different constellations of three ritual victims. We need not 
totally despair of such evidence in seeking to understand better the earliest 
Indo-Europeans since we are clearly discussing a series of domestic animals all 
known to the Proto-Indo-Europeans. We may hope to find some evidence for 
tripartite or triple sacrifice even if we cannot be secure of the precise beliefs that 
prompted the ritual behaviour. 

Horse sacrifice 

Some would maintain that the premier animal of Indo-European sacrifice and 
ritual was probably the horse. We have already seen how its embedment in 
Proto-Indo-European society lies not just in its lexical reconstruction but also 
in the proliferation of personal names which contain 'horse' as an element 
among the various Indo-European peoples. Furthermore, we witness the 
importance of the horse in Indo-European rituals and mythology. One of the 
most obvious examples is the recurrent depiction of twins such as the Indie 
Asvins 'horsemen', the Greek horsemen Castor and Pollux, the legendary 
Anglo-Saxon settlers of Britain, Horsa and Hengist (literally Horse and 
Stallion) or the Irish twins of Macha, born after she had completed a horse race. 
All of these attest the existence of Indo-European divine twins associated with 
or represented by horses. 

The major ritual enactment of a horse-centred myth is supported by 
evidence from ancient India and Rome and, more distantly, medieval Ireland. 
The Indie ritual is the asvamedha, probably the most spectacular of the ancient 
Indie ceremonies. It began in the spring under the direction of four priests, 
acting under the patronage of the king who was dedicating the sacrifice to the 
divine representatives of his warrior class. A prized stallion was selected as the 
victim and after rituals initiating the ceremony, the stallion was set free to 
wander for an entire year, 400 warriors trailing behind to ensure that the course 
of the stallion was neither interfered with nor that it had contact with mares. 



Ancilliary rituals took place throughout the year until the horse was returned 
for the final three day finale. This involved, among other things, the horse 
pulling the king's chariot, a large sacrifice of a variety of animals, and the 
smothering of the horse, after which the king's favourite wife 'co-habited' with 
the dead stallion under covers. The horse was then dismembered into three 
portions, each dedicated to deities who played out the canonical order of 
Dumezil's three 'functions'. 

The asvamedha bears comparison with the major Roman horse sacrifice 
which was known as the October Equus. Following a horse race on the ides of 
October, the right-sided horse of the team was dispatched by a spear and then 
dismembered, again in such a fashion as to indicate its 'functional' division into 
the three estates. As with the Indie ritual, the major recipient of the sacrifice was 
the warrior-god (Mars). In medieval Ireland, and through the admittedly 
somewhat jaundiced eyes of the Norman Geraldus, we read how in the 
inauguration of one of the tribal kings of Ulster, a mare was sacrificed and then 
dismembered. In a classic example of Ulster pragmatism, the pieces of horse 
flesh were then boiled in order to make a great broth in which the king 
subsequently bathed while devouring the morsels of meat. 

A detailed analysis of this and other material has led Jaan Puh vel to propose a 
Proto-Indo-European myth and ritual which involved the mating of a figure 
from the royal class with a horse from which ultimately sprung the famous 
equine divine twins. He offers some additional linguistic support for such a 
ritual in the very name of the Indie ceremony, the asvamedha. This derives from- 
the Proto-Indo-European *ekwo-meydho 'horse-drunk', attesting a ritual 
which included both a horse and drunkenness. This is quite comparable to the 
personal name Epomeduos which is found in ancient Gaul and appears to 
derive from *ekiPo-medu- 'horse-mead'. The modern English mead is trans- 
parently part of the same series that gives us Sanskrit madhu-, Greek methy, Old 
Church Slavonic medu, Lithuanian medus, Old Irish mid, and Tocharian B mit, 
all of which provide us with our word for the Proto-Indo-European alcoholic 
and ritual drink *medhu 'mead'. Hence, both the Indie and Celtic worlds still 
preserve the ancient Proto-Indo-European name of a horse-centred ceremony 
involving intoxication. 

The horse ritual warrants one more comment since it illustrates all too well 
how a comparison of myths may lead us along paths that appear to be 
contradicted by archaeological evidence. Both the asvamedha and October 
Equus clearly concern the sacrifice of a draught horse and in a striking instance 
of parallelism, both require that the horse in question excels on the right side of 
the chariot (cf. a Hittite ritual where the vehicle is drawn by a mule on the left 
side and a horse on the right). Clearly, this suggests that the horse is selected 
from a paired chariot team. But archaeological evidence indicates that the horse 
was not likely to have been employed in paired draught until the invention of 
the spoked wheel and chariot, which is normally dated after about 2500 bc and, 
consequently, some time after we would have assumed the disintegration of the 
Proto-Indo-European community. Indeed, the entire concept of horse-twins 



totally points to paired draught, while the archaeological evidence suggests that 
this should not be so at the time-depth we normally assign to Proto-Indo- 
European. Although cultural borrowing or parallel development may be 
suggested, this is a problem yet to be resolved. 

One final element of ritual associated with the horse sacrifice is the 
distribution of its anatomy after its death. In the asvamedha, for example, we 
have seen how the horse was butchered and offered to three different deities 
who can be assigned tripartite functions in Dumezilian fashion. As animal 
remains frequently accompany burials, we may hope that Indo-European 
rituals may shed some light on the patterns of offerings discovered in the 
archaeological record of the earliest Indo-Europeans. 

The cattle cycle 

We have already seen how lexical correspondences permit us to reconstruct 
proto-Indo-European expressions for 'to raid for cattle' and 'sacrifice of a 
hundred cattle'. At first glance we might regard these as the chance residue of 
the vocabulary concerning the secular (raiding) and sacred (sacrifice) disposi- 
tion of cattle in Proto-Indo-European society. But in an extensive examination 
of the role of cattle in both society and belief among the Indo-Iranians, and a 
number of peoples of East Africa, Bruce Lincoln suggests the paramount role of 
cattle in early Indo-European economy and religion. 

From mythological evidence primarily drawn from the Indians and Iranians, 
but also from the Greeks, Romans, Germans, Celts and Hittites, Lincoln 
reconstructs an Indo-European myth of the first cattle raid. This concerned a 
hero figure *Trito 'third' (Vedic Trita Aptya, Avestan Thraetaona Athwya, 
Greek Herakles, Norse Hymir, Hittite Hupasiya) who loses his cattle to a three- 
headed monster, normally a serpent, which at least in Indo-Iranian tradition is 
closely associated with local non-Indo-European populations. In a return 
encounter *Trito y with the assistance of the Indo-European warrior-god, 
defeats the three-headed monster and recovers his cattle. Lincoln suggests that 
this cattle-raiding myth served as 4 charter which both helped to define the role 

!!!!!!!!!!!!)!!!!! !!!!!!!! !!!!!!!? V.W.W \}\\\\ !!! !M !!t!?I!!!!! , !! , .*.!!!!I! T ! 1 1 !! TYY ! t ! nT 

j8 Bruce Lincoln suggests that the early Germanic Gallehus horn (c. AD 400) depicts a three- 
headed figure from the Indo-European cattle-raidtng myth. 



the celestial sovereigns 

79 The structure of the 
'cat tie-keepers' myth\ 

give cattle to 
the *arya people, 
who lose them to 

the priestly class, 
who sacrifice them to 

the *ddsa enemy, 
who steal them 
but are defeated by 

the warrior class, 
who recover them 
in raids and 
deliver them to 

of the warrior in Indo-European society (the proper activity of the warrior was 
cattle raiding), and sanctioned Aryan cattle raiding against foreigners who, 
according to the myth, had previously robbed the Aryans. 

Lincoln brings this warrior-class-centred myth into contrast with the myth 
of the first cattle sacrifice which served to underpin the position of the priest. 
This myth, about which we will have more to say below, involved the sacrifice 
of both a man and an ox (or bull) from w hose parts the w orld w as created. On a 
practical level this myth chartered the position of the priest who sacrificed 
victims to a sky-god who then bestowed both men and cattle on the kings or 
warriors of the Aryans in exchange. These were then expected to turn over 
cattle to the priest for sacrifice so that the cycle which secured the free flow of 
cattle through both human society and the cosmos was perpetuated. 

Lincoln argues that the striking similarities which he finds between the 
cattle-keeper's religion of the Indo-Iranians and East Africans is due to similar 
ecologies where possession of cattle defined the economic basis of society. Both 
cattle raiding to secure more of the principal commodity and cattle sacrifice to 
recompense and perhaps manipulate the deities were natural developments of 
such cattle-based societies. These two different activities encouraged the 
formalization of two separate classes - warriors and priests - whose own 
behaviour was patterned after the myths of the first cattle raid and first sacrifice. 
x\lthough Lincoln's study is primarily directed at the behaviour of the Indo- 
Iranians, his frequent recourse to general Indo-European mythology, espe- 
cially in the reconstruction of these mythic charters, suggests that the roots of 
the cattle-keeping religion and world view, with its attendant social ramifica- 
tions, might also be projected back to Proto-Indo-European society. 

Human sacrifice and punishment 

Human sacrifice is not a common occurrence among the rituals of the earliest 
Indo-European peoples, although there is hardly a group where some evidence 



for it cannot be found. In Germanic and Celtic tradition the evidence amounts 
to a reasonably well-supported pattern of The Threefold Death', wherein we 
can see human sacrifice or punishment applied in a clear trifunctional fashion. 
The manner of execution was carried out in a manner appropriate to the three 
Dumezilian functions. The ancient Gauls, for example, made offerings to three 
gods - Esus, Taranis and Teutates - by recourse to hanging, burning and 
drowning, respectively. This pattern is replicated in the pagan Germanic 
punishments of hanging, stabbing and drowning, each technique correlated to 
the crime for which the victim was convicted. The underlying scheme suggests 
that human sacrifice to a deity occupying a priestly or juridical role (or the death 
penalty for one who violated these particular interests) was death by hanging. A 
violation of the warrior code, or an offering to the god(s) of war, most 
appropriately awarded death either by burning or by the sword. Fertility deities 
were satisfied by drowned victims. Although the best evidence is primarily 
confined to the westernmost Indo-Europeans, there is some additional support 
to indicate that threefold sacrifices may have been more widespread. 

War of the Functions 

Certain striking parallels concerning the Roman account of the Sabine War, the 
Norse myth concerning the war between the Aesir and the Vanir, and the Indie 
epic Mahabharata have provided support for a Proto-Indo-European 'War of 
the Functions' from which some have drawn important conclusions about the 
formation of the Proto-Indo-European community. Basically, the parallels 
concern the presence of first- (magico-juridical) and second- (warrior) function 
representatives on the victorious side of a war that ultimately subdues and 
incorporates third function characters, for example, the Sabine women or the 
Norse Vanir. Indeed, the Iliad itself has also been examined in a similar light. 
The ultimate structure of the myth, then, is that the three estates of Proto-Indo- 
European society were fused only after a war between the first two against the 
third. From this mythic model, it has been suggested that the possible historical 
reality underlying the myth may be the conquest of settled agriculturalists by a 
non-sedentary community. This comes too close to one popular archaeological 
solution to the Indo-European homeland problem to pass without comment. 

The idea that there existed an historical reality behind the 'War of the 
Functions' is both highly speculative and unnecessary. We have already seen, 
for example, how the origin myth of the Scythians was constructed to serve as a 
social charter of behaviour and status within the Pontic region, that is, an 
Agathyrsi or a Geloni was subservient to a Scythian because his ancestors were 
incapable of stringing Herakles' bow, while Scythes, the ancestor of the 
Scythians, was successful. Similarly, if we admit a Proto-Indo-European 4 ^ar 
of the Functions', this need not reflect anything more than a reminder to the 
productive members of society that they remain subservient to both priests and * 
warriors, a situation divinely chartered by a mythical war which their ancestors 
lost but whose historical validity is no more secure than Herakles' bow. 



Dualism and Indo-European ideology 

We have already seen how Dumezil and his colleagues propose a pattern of 
dualism that cuts across the tripartite structure of Indo-European ideology. 
The first or sovereign function, for example, is expressed through paired gods 
(Varuna-Mitra, Jupiter-Dius Fidius, Odinn-Tyr) who are each respectively 
charged with the magico-religious and juridical-contractual aspects of 
rulership. The divine twins provide even more obvious evidence for dualism. 

The significance of twins in Indo-European mythology can be readily seen in 
the creation or foundation myths of the Indo-Europeans. The Proto-Indo- 
European *yem- 'twin' underlies the name of a god common to the Indo- 
Iranians (Indie Yama, Avestan Yimd) who becomes the progenitor of mankind. 
In a recent study, Jaan Puhvel argues that the underlying form for the name of 
Remus, the brother of Romulus in the story of the founding of Rome, was 
actually *iemus y the early Italic form of Proto-Indo-European *yemos 'twin'. In 
Norse mythology, mankind is formed from the remains of a giant whose name, 
Ymir, has also been derived by some from the Proto-Indo-European word for 
twin. Furthermore, Tacitus relates how the early Germans were the descen- 
dants of Mannus and Tuisto, the latter of which again means twin. Among the 
Celts we have the tale relating the foundation of Emhain Macha, the ancient 
capital of Ulster, which was explained by recourse to a myth in which Macha 
gave birth to emuin 'twins', again derived from Proto-Indo-European *yem-. 
Analysis of all these tales indicates that the Proto-Indo-Europeans believed that 
the progenitors of mankind were *Man (Indie Manu, German Mannus) and 
*Twin, the latter of which was sacrificed and carved up by his brother to 
produce mankind. To this Bruce Lincoln adds the coincidental sacrifice of a 
bovine integral to this myth in India, Iran, and among the Norse and Irish. 

We can go beyond the dualism expressed by twins to outright binary 
opposition as one of the underlying structures of Indo-European ideology. The 
most familiar example can be seen in how the Indo-Europeans treated the basic 
directions. As we have seen, the opposition between the Proto-Indo-European 
words for right and left also presents a systematic opposition between the 
concepts of propitious, healthy, strong, dexterous (Latin dexter^ Sanskrit 
daksina, Avestan dasina-, Lithuanian desine , Old Church Slavonic desn, Greek 
dexios, Old Irish dess, Albanian djathte, and so on, from Proto-Indo-European 
*deks-\ and the left which is unfavourable, unsound, weak, or, to use the Latin 
again, sinister. This opposition is also sexual since the right side or right hand is 
regularly associated with males and the left with females. Furthermore, the 
opposition also carries into the cardinal directions: the propitious south lay to 
the right (the Sanskrit and Irish words for right also mean south), while to the 
left lay the malevolent north, thus demonstrating that Proto-Indo-Europeans 
faced east to orient themselves. This right — left polarity is naturally not 
confined to the Indo-Europeans but can be found throughout the world. 
Nevertheless, securing this polarity to Proto-Indo-European society does 
provide the archaeologist examining the position and orientation of burials with 
another clue for tracking the course of the Indo-Europeans. 



The analysis of Indo-European ideological structure in terms of binary 
opposition is hardly removed from the structuralist approach of Claude Levi- 
Strauss who proposes a universal tendency to mediate between opposites. 
Bruce Lincoln has viewed the organization of Indo-Iranian ideology in a similar 
light, offering it as an alternative tool for understanding Indo-Iranian social 
theory. Here, for example, the *arya are contrasted with the aboriginal (and, 
from the Aryan perspective, inferior) *dasa. In turn, the *arya are subdivided 
into upper classes versus commoners, a system which Lincoln also finds in 
Caesar's account of early Celtic society. The upper classes are subdivided 
between sovereigns and warriors, while the sovereigns are composed of a binary 
opposition between priests and kings. In a somewhat similar vein, Einar 
Haugen has examined Norse ideology in terms of minimal oppositions. 

Some have sought an explanation for the ideological dualism in the social 
structure of Proto-Indo-European society. Tomas Gamkrelidze and 
Vyachislav Ivanov, following the work of other linguists, propose that the 
Proto-Indo-European system of marriage involved the exchange of women 
between two opposing moieties. This fundamental division of Proto-Indo- 
European society into two 'halves' could not help but engender an ideological 
response in Indo-European religion. Harkening back to earlier approaches to 
the mythological evidence, they propose an opposition between the two 
primary Indo-European deities: *dyeus pater, the god of the clear sky, charged 
with the maintenance of religious order, versus *perkuno - the god of storm, 
thunder and patron of war. More interestingly, they emphasize dual political 
leadership among the early Indo-Europeans. Citing Homer's account of the 
Achaian forces in the Iliad, they note how frequently the tribes listed are led by 
two rulers, while the dual kingship of ancient Sparta continued this tradition. 
Other evidence such as Horsa and Hengist leading the Anglo-Saxon settlement 
of Britain has already been mentioned. But is this meant to suggest dual 
kingship among the Proto-Indo-Europeans? We can no longer avoid the central 
problem of employing the evidence of comparative mythology to construct a 
picture of Proto-Indo-European society. 

Mythology and Reality 

Some of the critics of the 'new comparative mythology' harbour suspicions 
regarding how its proponents seem to tease out of any Indo-European 
document some evidence for tripartition. Others suggest that the three 
divisions of society proposed by Dumezil for the Indo-Europeans are so 
'natural' and generic to any society that they cannot be usefully employed as a 
diacritic for marking Indo-European culture. Some scholars engaged in the 
analysis of myth do not criticize Dumezil and his school; they simply ignore it as 
an irrelevancy to their own approach to the mythology of individual Indo- 
European peoples. But if one embraces the concept of tripartition, then it seems 
to offer unparalleled information for the archaeologist who wishes to correlate 
the concept of Indo-European with the archaeological record. 



The prehistorian is provided with a capsule description of Proto-Indo- 
European society divided into three major classes - priests, warriors and 
herder-cultivators. We may expect to see these classes marked by colour, totem, 
animal or any other form of culture-loaded symbolism. We may expect to 
discover the remains of animal sacrifices, especially that of the horse, or its ritual 
dismemberment. Triads of animals may occur in burials or other sacrificial 
contexts. Burials of males and females may exhibit variations in position and 
orientation similar to that indicated by the linguistic evidence. The most 
optimistic may even imagine that the Indo-European 'War of the Functions' 
underlies an historical reality, and search the archaeological record for the 
incorporation of an agricultural society into that of non-agriculturalists. How 
likely are we to be disappointed? 

There can be little doubt that the links between the reconstructed ideology 
and their expressions in material culture or behaviour of a prehistoric people 
may be far less than we hoped for in the last paragraph. Dumezil himself has 
insisted that his Indo-European civilization is one 'of the spirit', and that it need 
not be tied down to the real Proto-Indo-European world. Ideal worlds of 
myths, one may argue, are just that, and although they may be an expression of 
social realities, these need never take the corporeal forms required by the 
archaeologist. Squeezing priest burials out of the archaeological record of most 
Indo-European peoples, for example, has been a nearly impossible task. Any 
archaeologist examining the burial remains of the Celts in a La Tene cemetery 
may well wonder whatever happened to the facile generalizations concerning 
their social structure derived from myth and early ethnographers such as 
Caesar. The isolation of clear-cut classes or 'functions' is very difficult in Iron 
Age contexts when one might be more hopeful of discerning occupational 
classes than in the Eneolithic or Early Bronze Age. Perhaps most ironic is the 
fact that, even if archaeologists find it impossible to demonstrate the social 
structures predicted by mythologists, it will have absolutely no effect 
whatsoever on their continued publication of yet further examples of social 
tripartition and other aspects of Indo-European religion and ideology. 

If we must accept these difficulties then we must be honest about the utility 
of the new comparative mythology in elucidating Proto-Indo-European 
culture. Many of its practitioners would admit that the society which lies 
behind their reconstructions is an idealized one that need not be reflected in the 
cultural record, nor can it ever be effectively tested. Such an untestable 
hypothesis is, of course, no hypothesis at all, and its utility as an explanatory 
device is far better left to comparative mythologists who can play by different 
rules. They are not speaking the same language as the prehistorian. But how can 
we use arguments about the mythic reinforcement of social realities without 
assuming a relationship between social structure and myth? Here the resolution 
of this contradiction, perhaps more intuitive than logical, will be to hold the 
mythological evidence out for examination against the archaeological record, 
yet not make demands for proof higher than its own practitioners would 
willingly admit. 



The Indo-European Homeland Problem 

This Aryan family of speech was of Asiatic origin. 

A. H. Sayce, 1880 

This Aryan family of speech was of European origin. 

A. H. Sayce, 1890 

So far as my examination of the facts has gone it has 
led me to the conviction that it was in Asia Minor 
that the Indo-European languages developed. 

A. H. Sayce, 1927 

We begin our search for the homeland of the Indo-Europeans with the 
deceptively optimistic claim that it has already been located. For who would 
look further north than Lokomanya Tilak and Georg Biedenkapp who traced the 
earliest Aryans to the North Pole? Or who would venture a homeland further 
south than North Africa, further west than the Atlantic or further east than the 
shores of the Pacific, all of which have been seriously proposed as 'cradles' of the 
Indo-Europeans? This quest for the origins of the Indo-Europeans has all the 
fascination of an electric light in the open air on a summer night: it tends to 
attract every species of scholar or would-be savant who can take pen to hand. It 
also shows a remarkable ability to mesmerize even scholars of outstanding 
ability to wander far beyond the realm of reasonable speculation to provide yet 
another example of academic lunacy. It is sobering to recall that one of the 
greatest prehistorians of this century, V. Gordon Childe, dismissed his own 
researches into Indo-European origins as among the most childish things he 
wrote. It is no easy task to get one's bearings in a problem where most of the 
proposed solutions show a remarkable ability to be dismembered and securely 
entombed in one generation only to rise again to haunt later scholars. One does 
not ask 'where is the Indo-European homeland?' but rather 'where do they put 
it nowV Reflect for a moment that one of the most extensively argued solutions 
to the Indo-European problem at present has been advanced by the 
distinguished Soviet linguists Tomas Gamkrelidze and Vyachislav Ivanov who 
situate it either in or near Armenia, precisely in the same region that James 
Parsons set it nearly 230 years ago. The evidence and reasons for such a solution 
have naturally changed radically, yet the range of actual locations has not. So, 


So The modern 'consensus : A map of some of the solutions to the Indo-European homeland 
problem proposed since /g6o. 

lest one imagine that 230 years of research has carried us closer towards the 
truth, or perhaps that more negotiable concept of consensus, one need only 
glance at a map of some of the solutions to the Indo-European problem which 
have been proposed since i960. We hardly need any encouragement to pause 
briefly before embarking on our own quest to consider precisely what we are 
looking for. 

Defining the homeland 

Once we acknowledge that the historically attested Indo-European languages 
must derive from an earlier common or Proto-Indo-European language, logic 
also requires us to accept the existence of prehistoric communities which spoke 
that language. We may, of course, question the validity of the precise forms of 
our linguistic reconstructions and whether they represent exactly, or even 
approximately, the actual speech of the Proto-Indo-Europeans. We may also 
dismiss a w holly uniform dialectless proto-language as extremely improbable, 
and admit that there must have been considerable linguistic differences 
throughout the Proto-Indo-European speech community. But the actual 
existence of these Proto-Indo-European speakers, w ho must have left traces in 
the archaeological record of Eurasia, cannot be denied. 

When we attempt to reconstruct these Proto-Indo-Europeans we tend to 
telescope the cultural, linguistic and mythological similarities of the various 
historical Indo-European groups back into the prehistoric past. This tempts us 
to imagine the expansion of the Indo-Europeans from their original homeland 
much in the same fashion as a cosmologist surveys the origins of the universe 
from an initial 'Big Bang': running the cameras in reverse we follow the trails of 



energy and matter back to a singularity. But, for the Indo-Europeans, there can 
be nothing resembling a singularity. Only by assuming the preposterous notion 
that the Proto-Indo-European language originated simultaneously with human 
speech itself can we imagine it to have been anything other than a segment of the 
overall continuum of human speech in Eurasia. The Proto-Indo-European 
homeland is essentially the spatial expression of a vaguely defined temporal 
division of that linguistic continuum. Such a definition necessarily involves 
both the discipline of the historical linguist and the archaeologist, and each 
ignores the arguments of the other at his or her peril since the evidence for 
Proto-Indo-European speakers is a product of linguistics while the location of a 
prehistoric people is more properly the domain of the archaeologist. We have 
seen in Chapter Four that the evidence of the Indo-European vocabulary 
indicates that the Proto-Indo-Europeans should have existed broadly within 
the period 4500-2500 BC. There is both linguistic and archaeological evidence 
that can aid us in locating these Proto-Indo-Europeans within Eurasia at this 
time. There are also very serious problems of theory and method which will 
affect the validity of our solution. In the remainder of this chapter we will 
examine some of the approaches that have the greatest claim to our attention as 
well as some of the more general problems associated with the search for the 
Indo-European homeland. 

The neighbours of the Proto-Indo-Europeans 

A peculiar tendency among a number of nineteenth-century linguists was the 
strange desire to insulate the Indo-European homeland by great natural 
barriers, the Hindu Kush or the Himalayas proving particularly popular. Here 
primitive 'Aryans' were believed to have 'perfected' their language before 
bursting out over the rest of Eurasia. But if human experience be any guide, we 
must assume that during the fifth to the third millennia BC the Proto-Indo- 
Europeans were surrounded by other linguistic groups. In this section we will 
try to obtain a notional image of the prehistoric linguistic situation in Eurasia 
and to seek to locate the Proto-Indo-Europeans through their relations with 
their linguistic neighbours. 

First of all, we must try to imagine the Proto-Indo-Europeans occupying a 
territory among a large number of other linguistic groups. In Chapter Two 
we made a little rough estimation to postulate a territory of about 250,000- 
500,000 square kilometres for a major Bronze Age language, based on the 
historical disposition of the Near Eastern languages. If we were to take the main 
linguistic groupings of Europe today (Celtic, Germanic, Slavic, Finnic, 
Hungarian, Basque, and so on.), or to undertake calculations based on the major 
language groups attested for Iron Age Europe, the size of our individual 
linguistic territories would fall, on average, somewhere between 500,000- 
750,000 square kilometres. Furthermore, Sydney Lamb in his estimation of the 
number of linguistic entities in North America on the eve of European 
colonization reckoned about twenty-three language families which were 



represented by about 350 languages. If we apply these figures to the entire 
territory of North America, we would estimate that each of these language 
families would have occupied an area of about 1,000,000 square kilometres, 
while the individual languages would each have occupied about 65,000 square 
kilometres. Lamb estimated that, around 4000 bc, the number of languages in 
North America was probably of the order of 150-210, roughly half the number 
encountered historically. This would suggest an average size for each language 
somewhere in the range of 1 15,000 to 160,000 square kilometres. Similar rough 
estimates could be made from many more of the world's languages. It is, of 
course, not the precise figures that are important but the order of magnitude. 
Generally, before the emergence of major state languages we encounter most 
linguistic entities in the world occupying areas that range from the extremely 
small up to about 1 ,000,000 square kilometres. Consequently, we may postulate 
the size of the Proto-Indo-European homeland falling within the range of about 
250,000-1 ,000,000 square kilometres, a figure which is in fair accord with most 
solutions to the homeland problem as suggested by both linguists and 
archaeologists. A similar area is also suggested for the homelands of many other 
language families,, and where there are substantial deviations from such an 
average they tend to be downward rather than towards a yet larger area. 

We should remind ourselves why there seems to be an upper threshold to the 
size of a linguistic area, a 'maximum permissible area' as it has once been 
termed. Here we must recall that languages are always in the process of change 
and therefore, as the area of a given language grows in size, it will become 
increasingly difficult for all of its speakers to intercommunicate and change 
together along the same lines. Rather, there will be increased tendencies 
towards regionalization where linguistic change will follow different local paths 
of development. It is also highly probable that the economic system of the 
speakers of a given language will play a significant role in the size of area in 
which relative uniformity will be maintained. Mobile subsistence economies 
such as hunter-gatherers, or more certainly pastoral nomads, frequently retain 
linguistic uniformity over a wider area than is typically found among 
agriculturalists among whom long-term village settlement will probably 
promote regional developments. As the time-depth for Proto-Indo-European 
is set roughly to the Neolithic-Early Bronze Age, we may expect that there were 
a substantial number of different languages across Europe and neighbouring 
parts of Asia at this time. 

Our picture of a linguistically fragmented Eurasia 4500-2500 bc is also a 
natural consequence of accepting an average linguistic area of about 250,000- 
1,000,000 square kilometres. As with the North American evidence cited above, 
there remains today only a small number of language families attested in 
Europe whose ancestral proto-languages can be projected back into prehistory. 
We may conclude, therefore, that a considerable number of languages 
(potential language families) have totally disappeared sometime between 4500 
bc and the emergence of historical records in Europe. Applying our estimate of 
the average size of linguistic territory, there should have been something of the 



order of twenty to forty different languages occupying Europe at the same time 
as Proto-Indo-European. Now, it might strike one as improbable that so many 
languages or language families have disappeared without trace, but our earlier 
review of the historically attested disappearance of Hattic, Hurrian, Sumerian 
and Elamite in the Near East should have prepared us for just this sort of 
conclusion. Indeed, Iberian, Tartessian, Basque and Etruscan, remnant 
languages on the southwest periphery of Europe, suggest something of the 
magnitude of linguistic diversity that probably once prevailed over the rest of 
Europe. All of this emphasizes the incredible success that the Indo-Europeans 
achieved in expanding over territories many times greater than that from which 
they originated. This too is not without some analogic support. We know from 
the historical record that the Turkic language, confined in the sixth century AD 
to a region no larger than we would normally posit for Proto-Indo-European, 
virtually exploded over an area in excess of 2,500,000 square kilometres by the 
ninth century ad. Here, of course, the highly mobile character of Turkish 
society provided them with an advantage in their expansions and their ability to 
establish an exceptional linguistic uniformity over a gigantic region. 

Now we must push beyond the theoretical probability that the Proto-Indo- 
Europeans had many linguistic neighbours to see w hether it is possible to locate 
their homeland on the basis of their relationships with these neighbours. The 
remnant non-Indo-European languages of Western Europe, such as Basque, 
Iberian and Etruscan, offer us no evidence that would closely relate them to the 
Indo-European family, although one may always discern the occasional late 
loan word between Etruscan and Latin or Iberian and Celtic. Essentially, these 
languages offer us no compelling reason to seek an Indo-European homeland in 
Southwest Europe, nor has this area ever been a serious candidate. Of far 
greater importance in our search is the Finno-Ugric or Uralic languages. 

81 The historical expansion of the highly mobile Turks indicates how fast a language group 
wight spread. 

1 47 


82 Distribution of the Uralk languages and their probable homeland. 

The Finno-Ugric languages include Finnish, Estonian, Hungarian, and the 
Lappic, Permic and Volgaic languages. All of these are found to the west of the 
Urals, that is, in Europe itself, except for Vogul and Ostyak which occupy a 
wide area of the Ob drainage immediately east of the Urals. All of these Finno- 
Ugric languages may be taken as one major branch of a Uralic family whose 
other branch is the Samoyedic languages of northern Siberia. What is of 
greatest importance to the Indo-Europeanist is that these languages exhibit a 
convincing number of loan words borrowed from the Indo-European family 
when they were still largely undifferentiated in their own homeland. Knowing 
the location of the Finno-Ugric homeland permits us a rough approximation of 
the early Indo-Europeans. 

Some may naturally query using the Finno-Ugric languages as an anchor for 
the Proto-Indo-Europeans as seeking one unknown via another, but here we 
should remember that the Finno-Ugric homeland, despite the considerable 
amount of debate concerning its location, does not present the same magnitude 
of problems as the Indo-European homeland. The Finno-Ugric languages 
generally fill out a territory stretching from the northeast Baltic across Russia to 
east of the Urals (Hungarian was a medieval intruder into Eastern Europe from 
the Volga region and is most closely related to Vogul and Ostyak). This confines 
the search for the Finno-Ugric homeland to an area much smaller than the 
Indo-European one. In general, most proposed Finno-Ugric homelands tend 
to be confined to a strip about 1,000 kilometres long and 200-300 kilometres 
wide, stretching from the middle Volga east to the middle Ob, that is, an area of 
about 300,000 square kilometres. There are some who would seek the homeland 



further west as far as the Baltic, or to the southeast as far as the Aral Sea, but 
most solutions centre on the smaller area. Consensus, therefore, situates the 
earliest Finno-Ugrians in the forest zone of Eastern Europe-western Siberia 
among a variety of hunting-gathering-fishing cultures, many of whom may be 
embraced archaeologically under the Comb and Pit-marked Pottery tradition. 
It is now for us to see if it is possible to relate the Indo-Europeans to this region. 

The substantial number of loan words of clear Indo-European origin that are 
found across the Finno-Ugric languages suggest that they were borrowed when 
the latter was still a proto-language. The precise time at which these loans 
occurred is a delicate problem. It is generally acknowledged that many of the 
words were borrowed from the Indo-Iranian languages and not Proto-Indo- 
European, since the borrowed words have already undergone the characteristic 
changes expected of the Indo-Iranian languages. For example, we find that 
Finnish porsas and Votyak pars 'pig' are clearly derived from an Indo-Iranian 
*parsa - and not the Proto-Indo-European *porkos (cf. Latin porcus). This, and 
other examples such as Finnish sata 'hundred' which is comparable to Indie 
satam and Avestan satam (and not Proto-Indo European *kmtom y cf. Latin 
centum) all suggest that these loans were made sometime during the third or 
second millennia BC. While most would see these loans as unequivocally Proto- 
Indo-Iranian and hence notionally about 2000 ± 500 bc, a few have argued that 
they are distinctly Iranian loans although dating to approximately the same 
period. Consequently, we have reasonably solid evidence to maintain that about 
2000 bc Indo-Iranians were providing a series of lexical items, pertaining 
particularly to agriculture (such as pig, goat, grain, grass) and technology 
(hammer, awl, gold), to Finno-Ugric peoples situated roughly between the 
middle Volga and the Ob. This would place the Indo-Iranians to their 
immediate south (lower Volga-Ural) and helps confirm the location of the 
earliest Indo-Iranian tribes as suggested in Chapter Two. 

The evidence suggestive of a relationship between Indo-European and 
Finno-Ugric at a still earlier date is of greater interest. Here we are speaking of a 
series of possible loan words which would appear to have derived from Proto- 
Indo-European itself. In addition, there are also some striking similarities 
between the personal endings of the verb and some of the case endings of the 
noun between Proto-Indo-European and Proto-Uralic. Such similarities have 
suggested to some that the relationship between the two language families 
might not be two languages in contact but rather that they both derive from a 
still earlier common source which may be termed Proto-Indo-Uralic. Whatever 
way one reads the evidence, there is a substantial body of linguistic opinion 
which sees Proto-Indo-European and Proto-Uralic or Finno-Ugric as being 
closely related and consequently once adjacent to one another. With the Uralic 
languages occupying the northern forest zone, it is most probable that the 
Proto-Indo-European territory at least included a substantial portion of the 
forest-steppe or steppe zone of the Volga-Ural region. 

After Finno-Ugric, the possibility of Semitic-Indo-European relations has 
the greatest demand on our attention. Since well into the last century, and 



indeed still earlier, there have been frequent attempts to demonstrate parallels 
in vocabulary between the Indo-European languages and the Semitic languages 
of the Near East. The number of lexical comparisons, depending on one's 
source and to some extent imagination, is sizeable, and may approach 100-200 
among its more ardent proponents. Some of the comparisons invariably find a 
place in Indo-European handbooks such as Proto-Indo-European *(s)tauro 
and Proto-Semitic *tawru '(wild) 0 x', or Proto-Indo-European *septm and 
Proto-Semitic *sab y 'seven' (alone of the numerals). In the more extensive lists 
of comparisons, agricultural words constitute approximately a quarter of all the 
words allegedly shared by Indo-European and Semitic. 

Unlike comparisons between Indo-European and Finno-Ugric, the Semitic 
relations do not really have a general acceptance despite the fact that there are a 
number of most energetic supporters of genetic links between the two families. 
Certainly, any linguist running through some of the longer lists of comparisons 
encounters far too many cases of special pleading. Indeed, in a recent survey of 
the supposed Semitic — Proto-Indo-European loan words, especially those 
referring to agriculture and animals, Igor Diakonov has winnowed out almost 
all of the proposed comparisons except for goat, wild cattle and horn, all three of 
which he argued were probably derived from a common third source. Now 
there are, to be sure, some comparisons that will simply not go away such as 
Greek pelekeus, Indie parasu- 'axe' which is normally set beside Akkadian pilaqq 
'spindle, spike', itself a possible loan from Sumerian balag. This would appear 
to be a typical 'wander word ? moving along trade routes between various 
peoples from the Aegean to the Indus. Similar suggestions have been employed 
to account for possible similarity between Akkadian sarpu 'silver' and vaguely 
similar words in Germanic, Baltic and Slavic. But these, and occasionally other 
attempts to establish correlations between other Near Eastern languages, are 
quantitatively minute, qualitatively poor, and certainly do not reflect the 
unimpeachable pattern of evidence which we find with the agricultural and 
technological loan words that were borrowed from Indo-Iranian into Finno- 
Ugric. These indicate what intimate contacts between two language families 
should look like and this is not at all what we find in the Near East. There are, of 
course, some possible loans between individual Indo-European languages, as 
we have already seen between the Indo-Aryans and the Hurrians (Mitanni), or 
possibly Hittite and Hurrian. On the other hand, there is no convincing 
evidence that would set Proto-Indo-European in direct contact with the 
languages of the Near East although it is always possible that a few words may 
have passed between these two different families through intermediary 

A third family to be seriously proposed as possessing considerable similarity 
with Indo-European is Kartvelian. This is the major language group of the 
south Caucasus which is spoken by four million people, among whom the 
overwhelming majority speak the Georgian language. Although some lexical 
evidence has been cited (Tomas Gamkrelidze lists about twenty words), the 
primary case for Indo-European — Kartvelian relations is typological and 



concerns comparisons in the phonetic and grammatical systems of the two 
families. With such a case it is far easier to argue a very distant genetic link 
between the two families rather than some historical association played out 
during the period of Proto-Indo-European existence. Moreover, some scholars 
explain Kartvelian — Indo-European similarities by contact between the two, 
when the Anatolians came into long-standing contact with their Kartvelian 
neighbours to their east. 

Even more remote are the attempts to associate the early Indo-Europeans 
with the Altaic languages, especially Turkic, whose home is situated generally 
in the region of the Altai. Other than the occasional correspondence between 
Turkic and Tocharian, for example, Turkish Skuz 'ox' from Tocharian B okso y 
there is really no solid evidence to draw the Proto-Indo-Europeans so far to the 

To sum up this section, during the fifth to third millennia bc Proto-Indo- 
European was but one of a considerable number of languages spoken across 
Eurasia. It seems likely that the majority of the languages existing then have left 
no trace and now lie beyond our ability to recover. Of those languages to have 
survived into the historical record, those that comprise the Finno-Ugric family 
exhibit the strongest relationship with Indo-European and suggest that at some 
time they were both neighbours. That the Proto-Indo-Europeans shared a 
border with the Finno-Ugric speakers somewhere in the vicinity of the forest- 
steppe zone of south Russia appears to provide the most economical 
explanation of the evidence. But we would be overplaying our hand to imagine 
that it provided the only explanation since a contact relation or earlier genetic 
relation between Proto-Indo-European and Finno-Ugric could also have been 
played out directly west of the Finno-Ugric languages, as has been suggested by 
Oleg Trubachev. This would be an area that might include the northwest of the 
Soviet Union extending towards the Baltic and into Central Europe. But the 
attraction of a more southerly location for Indo-Europeans has the additional 
advantage of explaining any putative contacts with both the Caucasian and 
Near Eastern languages. 

Internal linguistic evidence 

It has long ago been argued that we may look within the structure, the 
distribution and the interrelationships of the Indo-European languages, to gain 
some insight as to where they were situated before their major dispersal or, at 
least, before they were earliest attested. The nature of this evidence and the 
approaches to it vary widely and we will proceed from some of the simpler forms 
of reasoning to a few of the more esoteric means of locating the Indo-European 

One of the most obvious approaches to our problem is to ask which candidate 
for a possible Indo-European homeland is most plausible from the geographical 
point of view. This basically requires us to set all other factors other than 
distance to zero and assume that the point of origin most conveniently located to 


the extremes of the total area of the Indo-Europeans provides the simplest 
solution. Naturally, this merely leads us to accept the centre of the language 
family's distribution as its most likely point of origin. When we apply such a 
rule to the distribution of the Uralic languages we would place a centre 
somewhere in the Urals, and for the Turkic languages we would seek its 
homeland somewhere in the vicinity of the Altai, locations which are at least 
congruent with the primary range of modern opinion. By the first century bc, a 
period late enough to encompass most known Indo-European languages, we 
find that they are distributed from approximately 10 degrees west longitude 
(the Celts of northwest Iberia and Ireland) to 90 degrees east longitude (Eastern 
Iranians, Tocharians). The north — south dispersion is much smaller lying 
between about 60 to 20 degrees latitude, that is, between the Germans and the 
I ndo- Aryans, and it is clear that the basic axis of Indo-European expansions 
should be east — west. Consequently, the centre point for the linear expansion 
would be approximately 40 degrees east longitude, the area of the Don river in 
south Russia. A homeland, therefore, that stretched somewhere between the 
Black and Aral Seas would be most convenient in terms of explaining the early 
geographical distribution of the Indo-European languages if only the factor of 
distance is considered. But a convenient explanation which ignores all the other 
variables is hardly a compelling one without further support. 

Any historical linguist would have good cause to argue that the dispersal of a 
language family is far more reliably measured in terms of languages than in 
kilometres, since it is the individual languages which form the constituent 
elements of the family. From this perspective there is something very lopsided 
about our selecting 40 degrees east longitude as the centre of dispersion since 
only the Armenian, Indo-Iranian and Tocharian languages are found east of it 
while all other Indo-European languages lie to its west. 

The western bias of the Indo-European languages was observed as long ago 
as the mid-nineteenth century when most scholars were automatically 
assuming that the homeland lay in Asia, the 'mother of nations'. It was at this 
time that the first serious proponent of a European homeland, the British 
ethnologist and philologist Robert Gordon Latham, argued that an Asian 
homeland violated the basic principles of natural science. When confronted 
with two branches of Indo-European - a European branch that occupied a vast 
area and which was extremely diverse versus the relatively homogeneous 
dialects of the Indo-Iranian group in Asia - one must naturally conclude that 
the latter were derived from the area of the former in the same way as one would 
derive an isolated species from the territory of its genus. Consequently, it was 
far more logical to imagine that the Indo-Iranian languages had moved away 
from the mass of other Indo-European languages rather than the converse, and 
to argue otherwise was to engage in the same type of absurdity as assuming that 
the Finno-Ugric languages originated in Hungary. 

Although Latham never occupied a central place in the development of 
historical linguistics, his line of argument is still widely utilized today. The 
principle implicit in his argument is that linguistic differentiation is a product 








83 The 'centre of gravity* of the Indo-European languages. 

of time, hence we may expect that where there is greater diversity the language 
family must have existed there for a greater time. Conversely, the appearance of 
relatively similar close-knit dialects suggests a recent occupation of a given 
territory. The idea of the 'centre of gravity' of a language family for determining 
the homeland area has been widely applied among the languages of the w orld, 
especially in North America but also in Africa. If we were to apply it here to the 
Indo-Europeans, we would find the area of greatest diversity to lie very roughly 
in Central-Eastern Europe where Italic, Venetic, Illyrian, Germanic, Baltic, 
Slavic, Thracian, Greek and possibly both Armenian and Phrygian took their 
more immediate origins. Such a solution pitches the homeland to about 20-30 
degrees longitude, considerably west of our earlier solution based solely on 
distance. It also raises two further issues or factors which themselves have 
served as routes to the Indo-European homeland. The first concerns the 
internal configuration of the Indo-European languages. While it may be quite 
attractive to dismiss the Indo-Iranian languages as a later offshoot of the 
European languages, the existence of the very 'un-Indo-Iranian' Tocharian 
languages forces us to consider other alternatives. Secondly, the centre of 
gravity principle, which implicitly depends on time as the major factor of 
linguistic diversification, ignores other, possibly more important, factors w hich 
might induce the linguistic fragmentation we encounter in Central-Eastern 
Europe. These are topics which we cannot afford to ignore. 

Earlier, when we examined the tree versus wave models of the Indo- 
European language, we noted the obvious tendency for various Indo-European 
languages to share common features w ith some, rather than others, of its sister 
languages. The most obvious division was the distinction between the centum 
and satem languages. In the nineteenth century, many linguists placed 
exceptional importance on the fact that Celtic, Germanic, Italic and Greek all 



maintained a Proto-Indo-European *k while eastern languages such as Baltic, 
Slavic, Armenian, and Indo-Iranian all revealed an s-sound. The west-east 
division suggested that the Proto-Indo-Europeans had divided like cell mitosis, 
and some even sought a major physical barrier to explain the sound shift, the 
Vistula river proving popular at one time. Naturally, the symmetry of this 
model was shattered when both the Anatolian and the Tocharian languages 
were also revealed to be centum languages. Some attempted to retain the 
traditional division by invoking migrations of heroic length from the west to 
explain the later positions of the Tocharians. The internal relationships of the 
Indo-European languages, therefore, does bear on their relative geographical 
positions and, consequently, on the Indo-European homeland itself. 

The internal configuration of the Indo-European language is essentially 
established by examining the phonetic, grammatical and lexical (vocabulary) 
similarities that are shared between each language. The fact that the centum 
languages retain the Proto-Indo-European while the satem languages 
change it to an s, is an example of a phonetic isogloss (similarity). An example of 
a grammatical isogloss is the augment, the prefacing of a Proto-Indo-European 
*e- to verbs to form a past tense, which is found in Greek, Armenian, Phrygian, 
and the Indo-Iranian languages (Greek ephere, Armenian eber, Sanskit abharat 
'he bore' from the root *bher- 'bear'), but nowhere else. Finally, there are 
isoglosses which involve vocabulary. For example, Italic, Celtic and Germanic 
all share the same word *netr 'snake' (Latin natrix y Old Irish nathir, Old English 
naeddre [Modern English adder] which is not found in any other Indo- 
European language). These isoglosses, which are numerous and varied, derive 
from either of two situations: they are examples of old Proto-Indo-European 
words or forms which have been retained only in some but have disappeared in 
other Indo-European languages; or they are innovations developed across 
adjacent dialects in a later stage of Indo-European after dialectal unity had 
collapsed. The word for snake cited above is a good example of the second 
situation since it cannot be ascribed to Proto-Indo-European but rather 
appears to be an innovation of the west Indo-European languages of Celtic, 
Italic and Germanic. Within historical linguistics there are considerable areas 
of disagreement over the status of many of the isoglosses yet there are also 
regions of broad, though by no means universal, agreement. Basically, they 
concern the nature of linguistic relationships experienced by the speakers of the 
various dialects between the period which we denote Proto-Indo-European and 
the emergence of each individual language. No solution to the Indo-European 
homeland problem can afford to ignore these configurations. 

The Anatolian languages appear to have preserved a number of archaic 
features, for example, laryngeals and heteroclitics, 29 and they lack a number of 
other characteristics such as grammatical gender, which is commonly found in 
the other Indo-European languages. Although the absence of some forms may 
certainly be attributed to attrition, since Anatolian was introduced into a 
territory densely settled by non-Indo-European speakers, there is still suffi- 
cient evidence to convince most linguists that the ancestors of the Anatolian 



languages separated from their other Indo-European relations at a relatively 
early time. It is precisely for this reason that the hypothesis of an early migration 
into Anatolia by Indo-European speakers is popular with many linguists. 

A central group of Indo-European languages comprised of Greek, Arme- 
nian, Iranian and Indie, would all appear to share some of the latest innovations 
in the development of the Indo-European languages, for example, the augment. 
This is historically explained by assuming that the ancestors of all of these 
languages were once geographically more closely associated than we find them 
in historical times. On the basis of our excursus into the individual origins of 
each of these groups we may suggest a chain of central Indo-European dialects 
stretching from the Balkans across the Black Sea to the east Caspian. 

There are other closely associated languages or configurations which 
conform well with the historical position of the various languages. For example, 
historical linguists recognize many shared similarities among the west Indo- 
European languages of Celtic, Italic and Germanic. Another recognizable chain 
of isoglosses links Germanic specifically with Baltic and Slavic while the latter 
two share a number of features with Iranian and Indie. All of these relations are 
geographically plausible and suggest that the relative position of most of the 
Indo-European languages has probably not altered greatly. By this is meant 
that we do not find compelling reasons to associate Celtic more closely with 
Iranian, for example, and hence have to devise complicated geographical 
hypotheses to explain their origins. In general, this indicates that the expansion 
of the Indo-European languages was broadly centrifugal from a more central 
homeland rather than linear from one of the extremes. 

Finally, we may glance at the temporal relationship between the various 
languages. We have already seen that the geographically more central 
languages, from Greek east to Indie, all appear to share late innovations. On the 
other hand, there are also several features that have traditionally been regarded 
as archaisms and which are preserved only in Celtic, Italic, Phrygian, Anatolian 
and Tocharian. The fact that the archaisms are situated on the periphery of the 
Indo-European world, while the innovations are found towards the centre, 
conforms with one of the well-known models of what is commonly known as 
linguistic geography: peripheral languages conserve while central languages 
innovate. To what extent this is a universally applicable norm is quite 
debatable, but this model does seem apt for the distribution of languages in the 
Indo-European family. In concrete terms it implies a central area composed of 
the ancestors of the Greek-Armenian-Iranian-Indic group, and a peripheral 
area of Indo-European dialects which both retained certain archaic features of 
Proto-Indo-European and avoided the later innovations of the central group. 
Temporally, this indicates that the ancestors of the peripheral languages had 
already dispersed before the central region had innovated. Geographically, a 
homeland which conforms to this model will have obvious attractions to the 
historical linguist. Again, then, the region between Central-Eastern Europe 
across the east Caspian provides a plausible homeland. Certainly, situating the 
Indo-European homeland to either the western or eastern extreme of the Indo- 



European dispersion would conflict with almost all of the evidence for the 
internal configuration of the Indo-European languages. 

Interference and substrates 

Our emphasis on how languages change naturally through time should not 
eclipse other factors responsible for the fragmentation of the Indo-European 
language family. While the expansion of the Indo-European languages requires 
human migration, it also must have involved the assimilation of a vast number 
of non-Indo-European peoples. Although the aboriginal populations would 
come to adopt the Indo-European tongue, the ultimate result might very well 
have been affected by their native language, its phonetic system, grammar and 
vocabulary. This is a phenomenon experienced by anyone who undertakes a 
second language, and we can hardly fail to hear the differences between English 
as it is articulated by a Frenchman, a Chinese or an Indian. The impact of the 
substrate language on the newly adopted one is known as interference and, since 
the mid-nineteenth century, linguists have often argued that the degree of 
interference provides a clue to the location of the Indo-European homeland. It 
is reasoned tha t where the Indo-Europeans tra veiled least from their homeland 
we should expect the least foreign interference and, consequently, the most 
conservative languages. Conversely, where we find the most radical deviations 
from Proto-Indo-European, we may assume strong interference from substrate 
populations whose territory should consequently be excluded from the 
homeland area. 

Theoretically, we should want to compare the proportion of each language's 
vocabulary, grammar and phonetic system with reconstructed Proto-Indo- 
European and then determine which languages seemed the most conservative 
and which had diverged the greatest. Unfortunately, we must ensure with such 
a procedure that we are comparing like with like. But how do we compare 
Hittite, for example, against Lithuanian, when the former was already extinct 
2,500 years before the latter was first attested? The later a language is first 
attested, the greater opportunity for it to have undergone changes stemming 
from its contacts with other languages after it has taken up its historical 
position. Greek, for example, which has a very long textual tradition, reveals 
approximately four times the number of inherited Indo-European words in its 
vocabulary as does Albanian which is first attested very late and, in the 
meanwhile, had borrowed heavily from its neighbours. Finally, there is no 
recognized means of measuring degrees of conservatism and weighing each 
change in phonetics, grammar or vocabulary against another. We must 
ultimately trust to impressions that cannot help but be subjective. 

Subjective or not, there are probably few linguists who would not admit that 
there is fairly clear evidence for substrate languages affecting some of the Indo- 
European languages. Even as early as the Vedic period, Dravidian loan words 
began to appear in Indie, while we have already review ed some of the evidence 
for non-Indo-European vocabulary in Anatolian and Greek. On a much 


9 Si 

qonui b uq p^JO puB uBqojBuy ui AjBjnqBDOA uBadojng-opuj-uou joj 
3DuapiA3 aqjjo auios paAvausj XpBajjB 3ABq ajiqM bipuj ui JBaddB 01 UBSaq 
spjoM uboj UBipiABjQ 'pouad sip^A 3l P SB ^I JE3 s ^ u^Ag s^BnSuBj UBadojng 

-OpUJ 3qj JO 3UIOS 5UU03JJB S32Btl3UB{ dJBJJSqitS JOJ 30U3piA3 JB3J3 X[JIBJ SI 3J3qj 

jBqj jiuipB jou pjnoM oqM sjsmSuq Avaj XjqBqojd sjb ajaqj 4 jou jo 9Ai:p3(qng 

3ApD3(qns aq jnq dpq jouubd jBqj suoissajdun oa jstu} XpjBUii:qn 
jsnm a^Y jaqjouB jsuibSb XjBjnqBDOA jo jbuiuibjS 'sopauoqd in ^SuBqD 

qDB9 SuiqSpAV pUB UISlJBAJ3SUO:> JO S33J<fop SuunSB3UI JO SUB3UI p3ZtuS033J 

ou si 3J3qj 'Xjjbuij sjnoqqSpu sji uiojj AnABaq paA\ojJoq pBq 'ajiqAvuBatu 

3qj Ul 4 pUB 3JBJ XJ3A p^JS^JB JSJIJ si qDiqM. UBIUBqjy saop SB XjBjnqBOOA 

sji ui spjoM uBadojng-opuj pajuaqui jo jaquinu aqi saum jnoj XpjBiuixojddB 
SJB3A3J 'uouipBJj jBmxaj Suoj Xj3a b SBq qoiqM 'ajduiBxa joj '^pajr) uopisod 
jBDUOjsiq sji dn u^bj SBq 11 mjv saStenSuBj J3qjo qjiM s;db^uod sji ujojj 
Suiuiuiajs saSuBqo auoftiapun 3ABq 01 ji joj Xjiunuoddo J3JB3j2 aqj 'pajsa^B 
isjg si aSBnSuBj b xs\v\ aqj^ ^pajswB jsjy sbav jwbj 3qj ajojaq sjbsX ooS'r 
iouijxo XpB^jjB sbav J3UU0J aqi uaqM 4 uBiuBnqjiq jsuibSb 'ajduiBxa joj l 3WUH 
ajBdiuoD op Moq jng a^q qjiM a^q SuuBduiOD 3JB 3M jBqj ajnpamid b 
qons qjiM 3jnsu3 jsnui 'XpiBunuojuQ jsajBajS 3qi paSjaAip pBq qDiqw puB 
oaubajosuo3 jsoui aqj paums saStenSuBj qaiqM auiiujojop uaqj puB uBadojng 
-opuj-ojojj p^pnjjsuoDQj qjiM uiajsAs oqauoqd puB jbuiuibjS 'AjBjnqBDOA 
spSBnSuBj qDB3 jo uoqjodojd aq* ajBduioo o^ jubav pjnoqs 3av 'XuBDuajo^q 


aqj uiojj p^pnpxa ^q Xpuanbasuoo pjnoqs Xjojuj^j ^soqAv suouBjndod 
aiBJjsqns uiojj 30U3J3jJ3jui Suojjs ^uinssB Xbui 3m 'uB^dojng-opu j-ojojj uiojj 


jsoui aqj 'Xpu3nb3SU03 'puB 3DU3J3JJ9JUI uSpjoj jsb^j 3qj J33dx3 pjiioqs 3M 
puBpuioq jpqj uiojj jsb^j pajpABJj suB^dojng-opu j 3qj 3J^qM jBqj psuosB3J si 
jj puBpuioq uB^dojng-opuj 3qj jo uoubdoj aqj oj 3np b s^piAOjd 33U3J3JJ^ui 
jo ^jS^p aqi jBqj panSjB uayo 3ABq sismSuq 'Xjnju^D qju^wuiu-piui aqj 

33UIS 4 pUB dOUdXZjxdlUl SB UM0U5{ SI 3U0 p^jdopB XjAV3U 3q^ UO 3^BnSuBJ 3JBJ3SqtlS 

aqj jo pBduit 3qj^ *UBipuj ub jo assuiq^ b 4 UBUiq3U3J j b Xq p^jBjnDUJB si \\ sb 
qsqSug u^^AVjaq ssDuajayip 3qj JB3q oi jibj XjpjBq ubd puB ^SbuSubj puosas 
b ss^Bjjapun oqAv 3uoaub Xq p33U3U3dx3 uou^uiouaqd b si siqjL XjBjnqBDOA 
puB jbuiuibjS 'ui^sXs oqauoqd sai '^BnSuBj saubu jpqj Xq p9j39j}B U33q 3ABq 
jpM Xj3a jq^iui jjnsoj ojBuiujn 3qj 'anSuoj UB^dojng-opui 3qi ^dopB o^ auioo 
pjnoM suopBjndod [BuiSuoqB ^qa qSnoqajv s^do^d uB^dojng-opuj-uou jo 
j^quinu jsba b jo uoub|iuiissb aqi paAjoAui ^ABq jsnui osjb ji 'uoubjSiui uBUinq 
ssjinb^j s^BnSuB] UB^dojng-opuj 3qi jo uoisuBdxa 3q; ajiqyW Xjiuibj sSBnSuBj 
uB^dojn^-opuj 3qj jo uoubjusuiSbjj 3qj joj ajqisuods^j sjojdbj jaqjo ssdipa 
jou pjnoqs ^uiii qSnojqi XjjBJmBU ^SuBqD ssSbiiSubi Moq uo siSBqdiua jiiq 

sa^Bj^sqns puB aDuajajjajuj 

s^BnSuBj uB^dojng-opuj aqa jo uopBjnSuuoo jbuj^ui 
^qj joj 3DU3piA3 aqj jo jjb jsouiyB qiiAv jDiguoD pjnoAv uoisj^dsip UB^dojng 

W3iaOHd (INV13WOH MV3dOHfl3-OaNI 3HX 


broader scale, the Celtic languages of both Britain and Ireland underwent an 
extremely brusque restructuring which has often been attributed to the 
influence of native non-Indo-European-speaking populations. These observa- 
tions, of course, only serve to exclude peripheral areas which we would have 
rejected as potential homelands on other grounds. There is for Northern 
Europe an extensive, bitter and thoroughly inconclusive controversy on 
whether the systematic change of all Proto-Indo-European consonants in 
Germanic should be attributed to non-Indo-European natives who spoke 
'broken' Indo-European. Further east we find far less discussion of substratal 
influences and a much greater acknowledgment of conservatism. This is 
especially seen in the case of Lithuanian which exhibits an astounding retention 
of the Proto-Indo-European forms as can be seen in the declension of the Indo- 
European word for 'wolf below: 





*wlk w os 




*wlk w e 




*wtk w om 




*wlk w osyo 




*wlk w od 




*w!k w oi 




*wlk w oi 




*wlk w o 



What is most striking is that Lithuanian shows roughly the same general 
retention of the Proto-Indo-European forms (naturally mitigated by minor 
sound shifts) as does Sanskrit, despite the fact that the latter language is attested 
nearly 3,000 years earlier than Lithuanian. This apparent archaism has 
mesmerized many linguists for over a century now and has led some to the 
conclusion that the Indo-European homeland must have lain in or near the 
Baltic. The case for a Baltic homeland has been augmented by a series of studies 
made by Wolfgang P. Schmid who has argued that the Baltic region even retains 
the Proto-Indo-European names for rivers. This hydronymic evidence we will 
pass over, since attempts to analyze river names in terms of Proto-Indo- 
European itself tend to be wildly subjective and seldom convince the majority 
of historical linguists. 30 Nevertheless, we are still left with the apparent 
conservatism of Lithuanian. Moreover, Vittore Pisani has observed that those 
languages west of the Baltic all show an abandonment of the Indo-European 
free accent 31 while Lithuanian and a number of the Slavic languages retain 
traces of it. And here we can observe that, although Slavic is not quite so 
conservative as Lithuanian, it still displays an extremely high retention of Indo- 
European noun forms. 

The evidence of Lithuanian, and to some extent Slavic, has predisposed 
many to seek the homeland in this region of Eastern Europe, or at least 
proximate to the Baltic and Slavic territories. It would be misleading to imagine 
that both of these branches of Indo-European did not show marked innovations 
as well as conservatism, and this is especially apparent in the verbs. 



Nevertheless, this cannot detract from the overall, subjective if you will, 
impression that the Indo-European languages of Eastern Europe have shown a 
stronger tendency to retain earlier Indo-European forms than have some of 
their neighbours. But this alone does not provide a secure solution to our 
problem. We have no more right to assume that interference is the prime cause 
of language change than the other factors upon which solutions have been 
constructed. Moreover, even if we were to attribute the conservative nature of 
Lithuanian to a lack of interference from non-Indo-European substrates, this 
need not indicate the absence of non-Indo-Europeans in the Baltic region but 
merely the effectiveness of intruding Indo-Europeans at assimilating a native 
population. Recall here the trivial impact of the Celtic languages of Britain on 
the development of English. Here some future linguist, ignorant of the evidence 
of both history and placenames, might conclude that England had always been 
occupied by Germanic-speaking peoples. 

While our excursus into the internal linguistic evidence cannot provide us 
with a conclusively demonstrated homeland, it does emphasize a recurrent 
pattern of support for a homeland which should lie between Central Europe 
and the east Caspian. 

Linguistic palaeontology 

One of the most widely recognized techniques for delimiting the Indo- 
European homeland is linguistic palaeontology, the same method employed in 
the reconstruction of the cultural vocabulary of the Proto-Indo-Europeans. 
When we apply it to the homeland problem, we compare the reconstructed 
Indo-European vocabulary against the archaeological and environmental 
record in order to determine what region in Eurasia best corresponds with the 
linguistic evidence. While there are the invariable problems of method 
associated with this approach, linguistic palaeontology does rest on a logical 
basis which can assist us in delimiting the Indo-European homeland. 32 

A starting point for all discussion must be the time-depth of our 
comparisons. All of the evidence so far reviewed indicates that the Proto-Indo- 
European language, as we reconstruct it, possessed a cultural vocabulary 
consistent with a date of roughly the fourth millennium bc. We may extend it 
somewhat in either direction, but not very much. Since we know that the 
Anatolian languages had already appeared by about 2000 bc and that the Indo- 
Iranians should have emerged by that time as well, any date more recent than 
2500 bc for Proto-Indo-European becomes increasingly implausible. There is, 
perhaps, a bit more latitude in extending the concept of Proto-Indo-European 
earlier than the fourth millennium BC, but anything much prior to 5000 bc 
would locate the Proto-Indo-Europeans temporally long before the earliest 
evidence for a number of the apparently later items - wagon, domestic horse, 
wool, yoke, plough - that are well established in the Proto-Indo-European 
vocabulary, plus others such as silver which are arguably Proto-Indo- 
European. Since one still comes across scholars asserting that the origins of the 



Indo-Europeans may well be traced to the Palaeolithic, we should pause to 
discuss our lower time limit for the concept of Proto-Indo-European. 

On brief reflection one could hardly dispute that the distant Palaeolithic 
ancestors of the Proto-Indo-Europeans possessed many words ancestral to 
those which later emerged in the Indo-European languages. We could hardly 
be surprised that Palaeolithic people, occupying the area of the homeland, had 
names for the trees, wild animals, implements, kinship categories, and natural 
features expected of any human society. On the other hand, we would be 
astounded if they possessed precisely the same words articulated in the same 
phonetic system and utilized with the same grammatical endings as their 
descendants many thousands of years later. This would violate all human 
experience. Rather, we might conjecture that a hunter speaking a distant 
ancestor of what later emerged as Proto-Indo-European may have called a cow 
**ngu in 1 0,000 b c, which later evolved into **ngwu by 7000 b c, then **gwuo by 
5000 bc to emerge in Proto-Indo-European as *g»>ous. Now, no form 
antecedent to the reconstructed Proto-Indo-European form can be termed 
Proto-Indo-European. This is not linguistic nit-picking but essential logic 
which, if ignored, leads swiftly to absurdity unless one is also content to thrust 
Proto-Englishmen into the Palaeolithic because the English language also 
retains reflexes of Proto-Indo-European vocabulary, for example, 'cow' from 
Proto-Indo-European *g» J ou$. Linguistically, the only way to proceed any 
further back in time than Proto-Indo-European is to assume the genetic 
relationship between Indo-European and another language family and attempt 
to reconstruct their common ancestor from their individual proto-languages. 
This has indeed been attempted on both a small scale, such as Proto-Indo- 
Uralic or Proto-Indo-Semitic, to vast constructs such as Nostratic which 
argues for a common ancestor for the Indo-European, Uralic, Altaic, 
Kartvelian, Hamito-Semitic and Dravidian languages. These controversial 
theories do not concern us here and we may merely emphasize that there were 
no more Proto-Indo-Europeans in the Palaeolithic than there were Proto- 
English or proto-Frenchmen. Our Proto-Indo-European is the slice of one 
particular strand of the linguistic continuum, falling about 4500-2500 bc. 

Once we have established the appropriate time-frame for our search, what 
items of the Proto-Indo-European vocabulary are most diagnostic for locating 
the homeland? In Chapter Four we reviewed the Proto-Indo-European 
environment but it did not offer many useful markers. The existence of a^word 
for mountain *g»>er(o) (Sanskrit girts, Greek deiros, Old Church Slavonic gora) 
have led some to assume that the Indo-Europeans must have originated in a 
mountainous landscape, candidates ranging from the Himalayas in the 
nineteenth century, to Armenia as recently suggested by several Soviet 
linguists. On the same line of reasoning, these latter exclude the plains of 
Northern Europe and the entire Pontic-Caspian region. Surely, one need not 
live on a mountain to have a name for it, and neither the North European plain 
nor the Pontic-Caspian steppe is so far distant from marked topographical relief 
that their occupants could have dispensed with a word for mountain. 



The arboreal vocabulary of Proto-Indo-European has played a much more 
significant role in the homeland problem. In general, most of the tree names 
that can be solidly attested for Proto-Indo-European are so ubiquitous that 
they cannot be used to delimit the homeland much further than we already 
have. Only the beech has provided more circumscribed limits for the homeland, 
on the traditional argument that it did not grow east of the Konigsberg-Odessa 
line and so excluded both a steppe and an Asiatic origin for the Indo- 
Europeans, but we have already seen that this word cannot bear the weight of 
such inferences. Even if we were to assume Proto-Indo-European status for it, 
despite the lack of a single iVsiatic reflex of the word and grave doubts about its 
original meaning, then the beech excludes Atlantic Europe, and that part of 
Europe east of a line from Moscow to Rostov on Don, but not the Caucasus 
where a variety of beech is well represented. On the other hand, the Volga and 
east Caspian do lie decidedly outside the distribution of the beech. 

Of the Indo-European wild fauna, almost all the reconstructed species have 
such broad distributions that they do not normally lend themselves to 
delimiting the homeland. The one major exception, at least from the 
perspective of the history of the homeland problem, is the Proto-Indo- 
European word *loksos which has frequently been taken to mean 'salmon' 
(Salmo salar). The salmon has a very restricted distribution in Europe and its 
imputation to Proto-Indo-European suggested that the homeland must have 
included the rivers which drain into the Baltic, hence Northern Europe. More 
recent and exhaustive examination of this word has, however, prompted 
Richard Diebold to reverse this conclusion. According to Diebold, the evidence 
indicates that the original referent was not 'salmon' but probably the 
anadromous 'salmon trout' {Salmo trutta) which is ubiquitous. In addition, a 
number of linguists have also argued for the reconstruction of the smaller 
speckled 'brook trout', thus providing two salmonids for Proto-Indo- 
European. Diebold presses the argument further by observing that none of the 
other salmonids - huchen, char, greylings - is attested for Proto-Indo- 
European and that their names were created as each individual Indo-European 
language came into contact with these new species. Now, there is only one area 

84 The 'beech line* and the 
distribution of the salmon, 
Salmo salar. 


in which the two salmonids attested for Proto-Indo-European are found to the 
exclusion of all others - the Pon tic-Caspian. Consequently, the fish that was 
once employed to dismiss the steppe region can be used to support it. To decide 
the homeland problem on such an argument involves us in the subtle issue of 
the admissability of negative evidence. 

To argue the location of the homeland from what items are not attested to 
Proto-Indo-European involves us in a silentio evidence. This has been 
frequently employed (and criticized) in the history of the problem. If the Indo- 
Europeans do not share a common word for amber, for example, how can one 
imagine that the homeland included the Baltic region where amber was of such 
importance? How can the homeland be situated in the Mediterranean or the 
Near East if we cannot reconstruct to Proto-Indo-European words for oil, 
olives, grapes, the ass, or many other environmental terms typical for this 
region? While the absence of a single word can be set aside as a matter of chance, 
can the absence of an entire series of words relating to a specific environment be 
so quickly dismissed? Or are all arguments based on negative evidence, to be 
dismissed as inconclusive no matter what their quantity or pattern? Linguists 
know all too well that, if we apply any form of strict scrutiny in determining the 
Proto-Indo-European status of a word, there are numerous examples of words 
which the Proto-Indo-Europeans must have had but which we cannot 
reconstruct, for example, we can reconstruct 'eye' and 'eyebrow' but not 
'eyelid'. Furthermore, we have already seen how one may reject tying a word 
and the homeland too close together when all one needs to assume is that the 
bulk of the Indo-Europeans moved outside the distribution of the particular 
item at an early time. As no one has been able to resolve this epistomological 
issue, the reader must decide how much worth should be placed on these 

A corollary to examining patterned absences in the proto-lexicon is the 
analysis of the shifts in meanings of reconstructible Proto-Indo-European 
words. A good example is provided in Proto-Indo-European *bhergo- 'birch'. 
This is one of the few tree names that can be strongly reconstructed to Proto- 
Indo-European. It denotes the birch in Indie (bhurja-), Iranian (Ossetic b'drz), 
Germanic (birch), Baltic (Latvian berzs) and Slavic (Russian bereza). However, 
its meaning has been shifted in Italic where Latin fraxinus refers to the ash, and 
the word is absent from Greek. Generally, linguists have argued that its shift or 
absence in Mediterranean languages is motivated by the fact that the birch is 
either absent or rare here, and hence when the Indo-Europeans first entered 
this region they either dropped the word from their vocabulary or shifted the 
meaning of the word to another species. In Paul Friedrich's extensive list of 
Proto-Indo-European (or early Indo-European) tree names, about half of the 
Greek reflexes of these words exhibit shifts in meaning, a pattern which is quite 
congruent with other evidence for the later Indo-Europeanization of Greece. 

If we move from the wild to the domestic fauna, then we find several items of 
interest. The most obvious and frequently discussed is the horse which is very 
well reconstructed to Proto-Indo-European and is normally taken to refer to 



the domestic horse. This is not where the matter ends, however, since we know 
that wild horses during the Proto-Indo-European period were also to be found 
over a broad area of Eurasia, including areas where we also know of domestic 
horses. Yet there is only one word for horse in Proto-Indo-European, and it 
does not distinguish between wild and domestic. The consequence of this is 
that while Proto-Indo-European *ekwos probably refers to the domestic horse 
it might well also have included the wild horse. 

During the relevant period, Eurasia can be divided into three broad 
categories with regard to the exploitation of the horse. The first category 
includes all of those areas where there appears to be a virtual absence of horse 
remains. This would mean most of the Near East, including Anatolia, with a 
very few exceptions which are normally attributed to the importations of horses 
from the steppe to the north. It also comprises the Balkans. The earliest horses 
in Greece do not appear prior to about 2000 bc which is consistent with a late 
Indo-Europeanizing of Greece. There is a little early evidence for horses on 
some east Balkan sites but this is generally interpreted as the westernmost 
extension of either wild horses or possibly contacts with domestic horse- 
breeders from the Pontic-Caspian. There is no evidence whatsoever for horses 
in the Neolithic period of the Carpathian Basin although very small numbers of 
horses do begin to appear, again from the east, during the fourth millennium 
bc. The other regions excluded from the primary range of the horse lie outside 
of serious homeland consideration, for example, the Apennine peninsula, 
southern Iberia, and Ireland, although they further emphasize the later 
movements of Indo-Europeans into these regions. If the horse, either wild or 
domestic, is employed as a major marker of the Proto-Indo-Europeans, then we 
would exclude most of Southeastern Europe from the homeland. 

The second category includes areas where horses appear to have survived the 
Late Pleistocene and continued to be exploited through the Neolithic, albeit as 
wild animals and as extremely marginal resources. Here we speak of the 
occasional presence of horse bones on sites of Central and Northern Europe. 
While the number of bones may not be great, they do occur at a sufficient time 
and distance from the major East European centre of horse breeding so that 
they cannot be immediately dismissed as merely the product of contacts with 
the Ukraine. Consequently, unless one demands that the Proto-Indo-European 
*ekwos referred only to the domestic horse, we cannot use this animal to exclude 
either the Neolithic cultures of the Danube or the TRB (Trichterbecher, that 
is, Funnel Beaker) culture of Northern Europe as possible Indo-European 

The third category is the actual centre of the natural range of wild herds of 
horses, specifically the tarpan, and the earliest-known centre of horse 
domestication. The main core of this region is to be found from the Dnieper 
river east to the Volga, and possibly on into Asia. It is extremely difficult to 
determine its eastern limits, but the presence of domestic horses in the 
Afanasievo culture of southern Siberia by the third millennium may suggest 
earlier predecessors. During the fourth millennium bc, the domestic horse 


appears to have expanded westwards and accounts for the increasing 
percentages of horse remains in the northwest Pontic and the earliest 
appearance of the horse in the Balkans and Carpathian basin. Thus, if one 
wishes to confine *ekn>os to the domestic horse, then linguistic palaeontology 
suggests that the border of the homeland should not lie much further west than 
the Black Sea. 

In evaluating the evidence from these three categories, we recognize that the 
Indo-European homeland may be most conveniently set to the Eurasian steppe, 
but we are under no compulsion to come solely to this conclusion. If the Proto- 
Indo-European *ekwos included also the wild horse, then there are no grounds 
for confidently excluding Central and Northern Europe as potential homeland 
territory. Still, it must be admitted that they offer slight support for the horse- 
centred society which we commonly project to the Proto-Indo-European 
period, and on balance the evidence of the horse better supports the steppe and 
forest-steppe. Finally, it is difficult not to see the absence of the horse in the 
Balkans as serious grounds for again challenging its claim to be included in the 

The earliest evidence for wheeled vehicles outside of territories either 
demonstrably non-Indo-European, for example, Sumer, or implausibly Proto- 
Indo-European - such as the Kuro-Araxes culture of south Transcaucasia 
where Hurro-Urartian languages appear - is to be found among a number of 
fourth-millennium bc cultures in Europe. These include the TRB culture of 
Northern Europe which reveals an acquaintance with wheeled vehicles by the 
mid-fourth millennium bc; the Late Copper Age Baden culture in the 
Carpathian Basin and other late Copper Age cultures of northern Italy which 
date from the latter half of the fourth millennium bc; and the Pontic-Caspian 
region where numerous remains of wheeled vehicles begin to emerge towards 
the end of the fourth millennium bc. Tomas Gamkrelidze and Vyachislav 
Ivanov, interestingly enough, have noted that one of our words associated with 
wheeled vehicles, Proto-Indo-European *k w ekHo- bears striking similarity to 
the words for vehicles in Sumerian gigir, Semitic *galgal- y and Kartvelian 
*grgar. With the putative origin of wheeled vehicles set variously to the Pontic- 
Caspian, Transcaucasia or to Sumer, we may be witnessing the original word 
for a wheeled vehicle in four different language families. Furthermore, as the 
Proto-Indo-European form is built on an Indo-European verbal root *k w el- 'to 
turn, to twist', it is unlikely that the Indo-Europeans borrowed their word from 
one of the other languages. This need not, of course, indicate that the Indo- 
Europeans invented wheeled vehicles, but it might suggest that they were in 
some form of contact relation with these Near Eastern languages in the fourth 
millennium bc. 

In general, the other solidly reconstructed technological items tend either to 
reinforce a border ranging from Northern Europe across the Pontic-Caspian, or 
to expand it considerably. The plough, for example, solidly attested to Proto- 
Indo-European, is known by way of ploughmarks in a number of fourth- 
millennium bc sites in Northern Europe extending from Britain to Poland. 



Copper is widely attested throughout Europe during the fourth millennium bc 
and cannot be employed as a useful geographic marker. Silver, somewhat more 
contentiously ascribed to the Proto-Indo-European vocabulary, does have a 
largely East European distribution throughout the later fourth millennium bc 
and, if accepted, would effectively exclude Northern Europe from the 
homeland area. But anyone conversant with the history of the homeland 
problem realizes that an argument dependent on a single item, particularly a 
linguistically contestable one, cannot hope to achieve universal agreement. 
Rather, we must take our lead from the general pattern of correlations between 
linguistic palaeontology and the archaeological record. 

A summary view of linguistic palaeontology engaged in evaluating the 
plausibility of an Indo-European homeland somewhere in Europe confines our 
attention primarily to the territory embracing the TRB culture of Northern 
Europe, the later Neolithic descendants of the Linear Ware culture of the 
Danubian drainage, the Eneolithic cultures of the Pontic-Caspian region and, 
perhaps, southern Siberia. The further one moves to the north, west or south of 
this broad band, the less plausible the territory becomes as a potential homeland 
congruent with the cultural vocabulary of the Indo-Europeans. Some areas, 
such as the Carpathian Basin, are merely transitional between what is 
acceptable and what is implausible. We have pushed the linguistic evidence 
about as far as we may; now it is the turn of the archaeologist. 


There have been over a century of archaeological solutions proposed for the 
Indo-European problem. The implicit assumption behind almost all of these is 
the belief that a linguistic entity can be located and its expansion traced in the 
prehistoric record. There has, unfortunately, been a suprising lack of concern 
for actually tying a prehistoric linguistic entity to the types of cultural debris 
encountered by the archaeologist. Anyone with the least familiarity with 
current archaeological literature recognizes that Gustav Kossinna's dictum of 
191 1, that sharply defined archaeological cultures invariably correspond with 
clearly marked ethnic groups, holds little attraction for today's archaeologist. 
Certainly we need not add to the litany of warnings that pots do not equal 
people. Nevertheless, while one may deny the necessity of assuming an 
invariable one-to-one correlation between an archaeological and a linguistic 
entity, it is equally perverse to assume that there can be no correlation between 
the two. There is sufficient evidence, for example, to indicate correlations 
between cultural trait lists and various linguistic groups in North America, 
which is hardly surprising since we might well expect that cultural traits are 
more easily shared by a people with the same language than between peoples 
excluded by linguistic barriers from easy intercourse and interaction. 33 Hence, 
archaeologists are seldom embarrassed at attributing the Celtic language to the 
bearers of La Tene remains, Germanic to the Jastorf culture, or various Bantu 
languages to certain African ceramic styles. The emphasis here must naturally 



be on inexact equations since one can always cite some exceptions - such as 
peoples sharing the same language but possessing radically different cultures, 
as in village versus Bedouin Arab; or different linguistic groups participating in 
cultures barely distinguishable from another, as among the linguistically 
diverse Pueblo Indian villages in the American Southwest. These exceptions 
are normally sufficient to dissuade most archaeologists from attempting to link 
a culture with a linguistic group unless the culture is proto-historic, like the 
Celts, and the gap between the archaeological and written records is not great. 
The linguistic identity of archaeological cultures more distant from the 
historical record may be thought to lie beyond reasonable inference. This is not, 
however, an option open to the archaeologist engaged in the Indo-European 
homeland problem, and we will have to follow the archaeological evidence as 
best we can. 

We may anticipate that any evaluation of an archaeological solution to the 
homeland problem will involve us in three major issues. The first is the nature 
of the archaeological entity selected for the homeland. We have already seen 
that there are some constraints on the area that we might select, although we 
must recall that at about 4000 bc the number and variety of potential homeland 
cultures in Eurasia is so great that the territory of the maximum permissable 
homeland area could normally accommodate at least several different cultures. 
This cannot help but lead to a certain vagueness no matter what our solution. 

Our inability to be precise is further enforced when we concern ourselves 
with tracing the actual migrations of the Indo-Europeans. Setting aside for the 
moment the enormous difficulty of tracing a folk migration in the archaeo- 
logical record, we must remember that we are also attempting to delimit the 
territory of a linguistic entity which we cannot date with the precision which we 
apply to an archaeological culture. We might, for example, equate the area of a 
culture dating from 4500 to 3500 bc with the homeland, when Proto-Indo- 
European may have not terminated until a millennium later, by which time it 
occupied a vastly larger area. Conversely, a culture dating from 3500 to 2500 bc 
might actually represent a stage in the linguistic continuum after the formation 
of individual Indo-European language groups. We are dealing with general 
trajectories rather than absolute borders when we attempt to apply the evidence 
of archaeology to the homeland problem. 

The more concrete problem here is naturally whether we can trace the 
movement of a prehistoric people in the archaeological record. For those who 
engage in such risks, there is basically a rough hierarchy of evidence that ranks 
burials as among the more significant evidence for migration and intrusion, 
closely followed by changes in architecture, ceramics and economy while other 
items of technology, especially metallurgy, are more often explained by means 
other than the expansion of ethnic groups. Much current modelling in 
archaeology tends to emphasize almost every possible means of explaining the 
spread of a new cultural manifestation other than actual migration. Exchange 
systems, prestige chains, peer-polity interactions, similar cultural evolution or 
internal structural reordering independent of external stimuli are frequently 



advanced against former models of culture changes that instinctively sought to 
introduce a new people with every new pot or burial. Such a predisposition 
against finding migrations in the archaeological record is primarily the luxury 
of prehistorians since any archaeologist operating within historical periods 
confronts so much evidence for both the historically attested and 
archaeologically evidenced movement of peoples that it would be ridiculous to 
dismiss them. Only a misguided (and I hope hypothetical) prehistorian would 
seek to explain Anglo-Saxon cemeteries in southern England as the result of a 
continental cult-package or the increased militarization of native British 
populations following the structural collapse of Romano-British society. 
Indeed, for the historical archaeologist the problem of correlation is often 
reversed - there is definite historical evidence for folk movement but precious 
little archaeological support. This raises one topic which simply cannot be 

Frequently, even when some evidence for migration is grudgingly acknow- 
ledged, it is claimed that the evidence is far too slight to account for the assumed 
linguistic changes. Those who pursue the matter further must engage in 
argument by decibel - it is the vehemence of the advocacy for or against 
invasion rather than an appeal to commonly accepted criteria of how much 
evidence is needed to validate or invalidate the case for a folk movement. If the 
intrusive origin of the English people is firmly reflected in the archaeological 
record, the Scots occupy the reverse position. Historical evidence relates how 
peoples in Ulster migrated to western Scotland in the fifth to sixth centuries ad. 
The linguistic evidence before this time clearly indicates that various Brittonic 
languages were spoken in Scotland prior to this colonization and after it we find 
the transplanted Gaelic language expanding through Scotland. Yet no 
archaeologist reviewing the sparse evidence for both Ulster and western 
Scotland would independently declare that there was sufficient evidence to 
substantiate a linguistic expansion from Ulster to Scotland. The same could be 
said of the introduction of Breton to Brittany from southern Britain, and many 
other historically attested but so far archaeologically unverified movements of 

The issue here, I would argue, is not so much the archaeological record but 
the misconceptions or at least preconceptions that many archaeologists bring to 
it, especially those who require unequivocal archaeological evidence for 
intrusions. Colin Renfrew has recently illustrated this attitude when he writes 
how archaeologists have now come to the 'realization that there have been far 
fewer wholesale migrations of people than once had been thought'. He is 
accurate in describing the general 'mood', at least of archaeologists in Western 
Europe, but he also exhibits the type of confusion archaeologists tend to share 
about linguistic problems. First, this new 'realization' has not, in my opinion, 
much to do with the actual number of migrations be they fewer or more; it 
concerns how archaeologists discuss the phenomenon of culture change in the 
archaeological record when there are no other historical or linguistic constraints 
on the material. Hence to suggest that the presence of a particular type of vessel, 

1 66 


the beaker to take a familiar example, and several other 'foreign' elements, were 
introduced into Britain by an immigrant population is now regarded as passe in 
many archaeological circles. Archaeologists would now seek other explanatory 
means for accounting for their presence. If, on the other hand, we had possessed 
later Neolithic texts in one language and found a new language expressed in 
'Beaker texts', then I am confident that the evidence previously dismissed as 
insufficient to support migration would be eloquently assisting us in 
determining the time of arrival and pattern of movements of these new people. 
This illustration is, I think, less hypothetical than it might seem. 

Witness for example how Renfrew himself treats the origins of the Insular 
Celts in his own study of Indo-European origins. The archaeological evidence 
for Iron Age intrusions is weighed insufficient to be interpreted as a migration, 
hence, we cannot explain the Insular Celtic languages by Late Bronze Age or 
Iron Age migrations into the British Isles. This throws us back onto the more 
archaeologically visible introduction of the farming economy into Britain and 
Ireland in the late fifth millennium bc. But when we come to later Celtic 
movements such as those into Scotland from Ireland during the first 
millennium ad a different model is introduced - elite dominance, suggesting a 
small number of people at the apex of a hierarchy who bring about a linguistic 
transformation. Why? Both Scotland and Ireland are of similar size; both come 
to adopt the essentially same Gaelic language. The archaeological evidence for 
intrusions into Ireland during the Late Bronze Age or Iron Age is certainly no 
poorer than that for movements from Ireland to Scotland in the first 
millennium ad. But linguistic and historical evidence constrains Renfrew to 
accept the later migration (without a shred of convincing archaeological 
evidence) and yet he rejects any Bronze or Iron Age migration (elites or 
otherwise?) into Ireland. I do not dispute that the dominance of an Irish elite in 
western Scotland probably led to the widescale adoption of Gaelic there, but I 
do dispute that any archaeologist can hope to pronounce on what linguistic 
displacements did not happen in the prehistoric record. 

I believe we must discern between two epistomological phenomena: 
evidence for putative migrations generated entirely by a closed interpretation of 
the archaeological record; and the issue of migration generated by a reading of 
the linguistic or historical evidence. In an ideal world there should be no 
distinction between the two and any archaeologist should be able to establish 
the validity of a particular invasion hypothesis independent of whether there is 
historical or linguistic evidence for it. My experience with the literature 
concerning both prehistoric and historic migrations convinces me that this is 
not yet the world we live in, and that archaeologists play by different rules 
depending on the nature of the problems confronting them. 

This whole problem of distinguishing linguistic intruders in the archaeo- 
logical record is aggravated by the fact that our ability to provide the necessary 
precise dating for the evidence of population movements lessens as we proceed 
into prehistory. A Dark Age migration that is at least apparent is normally based 
on data such as a series of well-defined cemeteries extending over one to several 



centuries and apparently intrusive into the region we find them. Although 
problems most certainly exist, archaeologists can at least point to intrusive 
Anglo-Saxon, Longbard, Gepid, Avar, or Sarmatian cemeteries. In much 
earlier periods, however, we neither have such abundant data nor can we order 
it within the fine tuning of a century, hence any intrusion that has resulted in the 
almost predictable adoption by the newcomers of elements of the native culture 
will seldom, if ever, be particularly obvious. Returning to the spread of the 
Athapascans in North America, we find some of them practising a sub-Arctic 
culture in their Alaska-Canadian homeland; others who migrated south into the 
Pacific Northwest were assimilated into the classic wealth and prestige- 
obsessed cultures of their neighbours; and those who pressed further south, 
adapting to both the new environment and acculturating towards the native 
occupants of the American Southwest. The language persisted but the cultural 
evidence of its speakers was shed each time it came into a new area. It is a 
fortunate prehistorian who recovers an intrusive population before it begins to 
assimilate towards the native culture. 

The lessons of all this are several. The archaeologist can identify what appear 
to be discontinuities in the archaeological record. He may interpret them as 
evidence of intrusions, or, as Renfrew remarks, reclassify what were once taken 
to be intrusions into something else. So long as there is no compulsion from 
linguistic or historical evidence, this is a game that archaeologists are best left to 
play alone. There is, I believe, a corollary to this: archaeologists would be better 
off not wasting their time assigning unnecessary linguistic identities to their 
putative intrusions. Arguments that the 'Beaker folk' may have introduced an 
Indo-European language into the British Isles (though not one that evolved 
into anything for which we have evidence), or that the appearance of the royal 
tombs at Alaca Hiiyuk in central Anatolia were the result of Indo-Europeans 
(but not the ancestors of the Hittites), seem to me to be idle. Dealing with the 
real problems of language intrusion is difficult enough without speculating 
about the spread of unattested languages. The archaeologist may also assert that 
there is no obvious evidence for an intrusion or a discontinuity at a particular 
time, though he or she cannot pretend to control all the possible data and 
exclude the possibility of a linguistic intrusion. The most positive pronounce- 
ment that can be made is that both archaeological and linguistic or historical 
evidence seem to be congruent with one another in suggesting linguistic 
intrusion; the most negative is that the evidence from the various disciplines do 
not seem to form a congruent picture. 

Finally, there is another issue that is almost invariably ignored yet which is of 
equal consequence as the previous one. Even if we feel that the evidence does 
admit of a folk movement, how do we assess that it effected a linguistic 
replacement? As we have seen, intrusive populations may well find themselves 
assimilated by their hosts rather than the reverse, especially when migrations 
will often involve a smaller population intruding into the territory of a larger. It 
is perhaps here that the analysis of the changing structures within society is 
most productive and we will seek to explore this further in Chapter Eight. 


i j The face of an ancient German. The famous Tolluml man, 
recovered from a Danish hog, i /lust rates the appearance of the Iron Age 
German. The rope around his neck suggests that he met his death through 
punishment or sacrifice. 

r8 (above) Cast bronze helmet plates from Torslunda in Sweden 
depict hard-pressed Scandinavians and monsters. The plates date to the sixth 
or seventh century ad. 

rg (right) Detail of tiro scenes of Iron Age life in northern Italy 
f rom the Certosa sit nla f rom Bologna. The upper register depicts preparations 
for a feast while the lower illustrates either a contest or duet involving two 


20 Scene from the Gundestrup cauldron. The plate illustrates troops 
of infantry and cavalry wearing Celtic regalia and blowing the typical Celtic 
horn, the carnyx. The scene on the left is generally interpreted as a ceremonial 

21 (above) Burial from Khvalynsk cemetery, Volga region. 
Accompanying ornaments include shell beads and copper rings imported from 

the Balkans. 

22 (below) Typical communal burial pit of the Dnieper-Donets 
culture from the site of Xikolskoye. 

23 ( (op left ) Burial from the Sezzhee cemetery in the middle 
I / alga region. The individual was accompanied by a stone axe, knife, 

head necklace and zoomorphic hone plates. 

24 (top right) Reconstruction of the robust Cro-magnon 
features of a Dnieper-Donets male from Nikolskoye. 

25 (above) Reconstruction of a Dnieper-Donets female. 

26 ( left ) Reconstruction of a male from the Sredny Stog site 
of Alexandria. For those who hold to a Pontic-Caspian homeland, this 
represents, in general, the physical type of the earliest Indo-Eiiropeans. 

27 r/zt' Kemosovka 
stele, discovered in rgyj, provides one of 
the finest examples of Eneolithic art of the 
Pontic region. It reveals a moustached 
figure with club, dagger and three axes 
above the belt, while below it are depicted a 
large rectangle and two animals, identified 

as horses. 

28 (below) Clay model of a 
four-wheeled wagon from the Baden ceme- 
tery of Budakalasz. This provides some of 
the earliest evidence for wheeled vehicles in 
Danubian Europe in the late fourth and 

early third millennia BC. 


Throughout this chapter all of the evidence reviewed so far predisposes us to 
assume that the Proto-Indo-Europeans were situated, broadly speaking, in a 
territory extending from Central or Northern Europe on the west across the 
Pontic-Caspian and possibly into southern Siberia. We see little reason to 
support an Indo-European homeland that included either Western Europe, 
Mediterranean Europe, the Balkans, or the Near East. Western Europe, and 
certainly the southern Mediterranean, have never really been considered 
serious candidates for the homeland. The same cannot be said, however, for 
Southeast Europe (the Balkans and Greece), since some linguists and 
archaeologists have proposed that the Proto-Indo-Europeans might be placed 
in this territory. We need to take a closer look at the case for and against 
Southeast Europe. 

In a world where the archaeological record refuses to shout out evidence for 
massive Indo-European migrations, one might well sympathize with Colin 
Renfrew when he retreats to the expansion of the Early Neolithic in Europe as 
the only 'basic and widespread cultural and economic change' that might 
account for the spread of the Indo-European languages. Here, Renfrew is 
following the conventional wisdom that maintains that Neolithic colonists from 
Western Asia, probably Anatolia, crossed into Greece during the seventh 
millennium bc where they established Europe's earliest Neolithic communi- 
ties. During the course of generations, the descendants of these colonists 
maintained a 'wave of advance' gradually carrying the new subsistence 
economy both northwards through the Balkans and westwards from Greece 
through the central and western Mediterranean. Further advances continued 
northwards along the Danube drainage to the Atlantic and Baltic, while others 
pushed eastwards around the northwest shores of the Black Sea where they 
adopted pastoral economies and spread further east across the steppe. The 
superior productive capacity of these early agriculturalists ensured their 
success at assimilating the local Mesolithic communities and the spread of their 
own (Indo-European according to Professor Renfrew) language(s) from 
Anatolia to the peripheries of Europe. The differentiation and internal 
relationships of the Indo-European languages were then established according 
to the 'wave model' of linguistic change. 

In establishing the mechanism of linguistic expansion, a 'wave of advance' of 
farming communities progressively colonizing new territories over a period of 
generations, Colin Renfrew accepts a traditional model for the spread of the 
Neolithic economy. This has been recently challenged by some revisionists 
who, like Graeme Barker, argue that 'the various systems of initial farming 
which we can discern were developed by the indigenous population rather than 
by newcomers'. This position is rather extreme, and seems to flourish among 
only a handful of British universities. On the other hand, most archaeologists 
specializing in the origins of the European Neolithic would not accept that the 
expansion of the farming economy could be entirely attributed to a wave of 
advancing farmers whose own origins lay ultimately in Anatolia. Rather, we 
distinguish in the archaeological record areas where a strong case can be made 



for actual colonization from Asia, for example, Southeast Europe (Greece and 
the Balkans), probably the Linear Ware territory of Danubian Europe, and 
other areas where the evidence of continuity from the preceding Mesolithic 
suggests the local adoption of the new economy by indigenous populations. 
This is especially seen along the peripheries of continental Europe, in the 
western Mediterranean, in Northern Europe and in the forest, forest-steppe 
and steppe zones of Eastern Europe. 

If we accept the traditional argument that farming communities did cross 
from Asia to colonize Southeastern Europe, then it would indeed be difficult to 
resist the notion that they served as vectors for spreading new languages into 
Europe. The issue is whether the new languages were Indo-European, as Colin 
Renfrew has proposed. Given all the evidence reviewed so far, I find this one of 
the least likely hypotheses. Since Renfrew's book is admittedly speculative, I do 
not intend to dwell on those points which seem to be particularly implausible, 
for example, early Neolithic Proto-Indo-Iranians in the Zagros and Indus, 
Troto-Irish' in Ireland about 4500 bc, and so on, since they are peripheral to 
the main thrust of his argument. Rather, I will examine the core of Renfrew's 
solution to the Indo-European homeland problem as this also has a bearing on 
the alternative solutions. 

Renfrew draws the Indo-Europeans from the earliest Neolithic communities 
of Anatolia, including eastern Anatolia, broadly the region from £atal Hiiyiik to 
Cayonii. As we have seen before, with the emergence of the earliest written 
sources, this region not only exhibits Indo-European populations (Hittite, 
Luwian and Palaic), but also at least two apparently unrelated non-Indo- 
European groups: Hattic and Hurrian. Thus, part of the area designated as the 
Indo-European homeland emerges as non-Indo-European with our first 
historical records. The only way I can account for this within the framework of 
Renfrew's model is to assume that later non-Indo-European languages spread 
into the area of the former homeland, or to assume that two or more language 
families grew up in the same area and at the same time. There is little to be said 
in favour of either idea. I am puzzled by a solution that propels Indo-Europeans 
out of their Anatolian homeland such that they traverse over 3,000 kilometres to 
arrive in Ireland in the course of two millennia, yet require 5,000 years to 
advance the 100 or 200 kilometres (if that much) east to Armenia? When the 
Armenians occupied the former Urartian kingdom about Lake Van, was this a 
reoccupation of the Indo-Europeans' former territory, or the furthest eastern 
expansion of the Indo-Europeans from their Anatolian home? Since neither 
Phrygian nor Armenian can be said to have evolved from Anatolian, and their 
presence in their respective locations surely requires some form of migration, it 
seems that in order to sustain Renfrew's model of Indo-European origins we 
have to propose some form of reflux movement and later migrations which are, 
of course, quite contrary to his attempt to propose a simple archaeologically 
visible migration of Indo-Europeans. 

An Indo-European homeland which was confined to only south-central and 
western Anatolia (and hence avoided having to explain Hurrians and Hatti in 



the Indo-European homeland) could be regarded as a geographical adjustment 
which might avoid some of the former criticism, although one might note that it 
requires us to derive the Indo-Europeans from only a handful of sites. But 
problems still exist since we simply cannot ignore the substantial proportion of 
the placenames of Anatolia that are unanalyzable as either Indo-European or 
within the framework of the Anatolian dialects, for example, Parnassos (in both 
Anatolia and Greece) with a root parna 'house' attested in Hurrian purni and 
Egyptian pr. In short, there is a case to be made for non-Indo-European 
substrates across all Anatolia and on into Greece, and I would have thought it 
far more plausible to associate these with the spread of the Neolithic economy 
from this region rather than what are almost universally taken to be intrusive 

The time-depth for an expansion of the Indo-Europeans in the early 
Neolithic about 6500 bc is wholly incongruent with our reconstruction of the 
Indo-European vocabulary. Renfrew attempts to address this problem either 
by denying the utility of linguistic palaeontology or by attempting to defend his 
proposed homeland against its findings. For example, the presence of copper in 
early Neolithic Anatolia indicates that his 'homeland' cannot be excluded on 
the grounds that the Proto-Indo-Europeans had a word for copper. Neverthe- 
less, as we have seen before, terminology for wheeled vehicles is so abundant 
and deeply embedded in the Indo-European languages that we must accept 
their ascription to Proto-Indo-European if the comparative method means 
anything. Similarly, the horse, which is not attested in any form in Anatolia 
until the fourth millennium BC, and in Greece until the third, and is again 
thoroughly embedded in the reconstructed vocabulary as well as in Indo- 
European ritual, makes a seventh-millennium bc homeland in Anatolia or 
Greece quite unacceptable. If Andrew Sherratt is correct in placing the 
development of his 'Secondary Products Revolution' in the Late Neolithic, 
then words for yoke, plough, wool, and so on, all solidly attested to Proto-Indo- 
European, would also render an association between the earliest Neolithic and 
the Indo-Europeans extremely unlikely; the same might be said of Proto-Indo- 
European silver. While most wild animal terms are rather ubiquitous in 
Europe, Anatolia and Greece do seem to lie beyond the range of the beaver, 
another word which can be ascribed to Proto-Indo-European. Even the social 
and kinship terms which indicate at least that the Proto-Indo-Europeans were 
patrilineal and 'patriarchal' (and possessed some form of warrior sodality) do 
not seem to fit the agricultural societies we find in the early Neolithic cultures of 
Greece and the Balkans. 

An Indo-European homeland in Anatolia makes it impossible to explain the 
structural and lexical similarities between Proto-Indo-European and Uralic, 
the linguistic family that would appear to be most closely related to Indo- 
European. Rather, we are required by Renfrew to seek the emergence of Indo- 
European in a milieu of non-Indo-European languages, such as Hattic and 
Hurrian, which are structurally worlds apart from Indo-European and provide 
no grounds to suggest anything other than late contacts. Given that the time- 



depth suggested by Renfrew for Proto-Indo-European is also broadly 
contemporary with the emergence of the early Neolithic societies in Southwest 
Asia, I think we would have the right to expect far greater similarities between 
Proto-Indo-European and the other languages of this region if this were the 
home of the Indo-Europeans. 

The evidence for a non-Indo-European substrate in Greece, shown in 
placenames, personal names, tribal names, and cultural terms - some post- 
dating the Neolithic - seems a reasonable supposition despite all of the 
problems in sorting out the various layers. Moreover, we have other evidence 
such as shifts in tree names which make a circumstantial case for assuming an 
expansion of Indo-Europeans into the Mediterranean and not out of it. 

Concerning the origin of the Indo-Europeans of Asia, Renfrew must press 
his wave of advance around the northwest shores of the Black Sea to explain the 
origin of Indo-Iranian pastoralists of the Pontic-Caspian steppe and beyond. 
We will see in the next chapter that such a model is specifically rejected by most 
if not all Soviet archaeologists engaged in the study of this region. Unlike other 
peripheral areas where we might propose later expansions from Central Europe 
during the Bronze or Iron Age to protect Renfrew's hypothesis, there is no body 
of evidence to suggest anything other than the gradual acculturation of steppe 
populations well beyond the frontier of the 'wave of advance' colonists. Failure 
to account for the origins of the Indo-Iranians (and Tocharians) invalidates any 
putative solution to the homeland problem. 

An Anatolian homeland is at variance with the time-depth for the 
fragmentation of the Indo-European dialects, based on linguistic criteria. 
According to Renfrew's theory, the language ancestral to Greek was already in 
Greece by about 6500 bc. Indo-Iranian, following the less implausible of 
Renfrew's two (implausible) hypotheses, 34 was spoken by descendants of the 
earliest farming communities who pushed around the Black Sea by the fourth 
millennium bc and then adopted a pastoral economy and expanded across the 
steppe. This would indicate that, when we first encounter these languages, they 
have been separated for over 5,000 years despite the fact that they share what are 
generally regarded as numerous late isoglosses. The temporal separation 
between them, according to Renfrew's theory, seems wholly incongruent with 
what we understand of the Indo-European dialects, and we can only achieve 
this (lesser?) implausibility by selecting an archaeological interpretation of the 
origins of the steppe pastoralists that seems contradicted by the archaeological 
evidence. There are other examples of how Renfrew's theory disregards any 
notion of the internal relationships of the Indo-European languages. He derives 
Italic from a seventh- or sixth-millennium Neolithic expansion from Greece 
across to Italy, yet we know that Italic and Greek are divided by far more 
isogloss bundles than join them. In short, the whole question of the time-depth 
involved with his expansions is at variance with almost all current thinking on 
the problem. 

One could compile more objections, such as, if the Indo-Europeans were in 
Crete since 6000 bc, why can't we read Linear A? Why does it appear to have 



been devised for a language whose syllabic structure is radically different from 
the structure of Indo-European words? I would argue that the cumulative 
weight of these arguments indicates that any attempt to tie the initial Neolithic 
colonization of Europe to the spread of the earliest Indo-Europeans is really not 
congruent with either the linguistic or archaeological evidence and, indeed, 
does not even provide the economy of explanation which should have been one 
of its major attractions. Anatolia is the wrong place at the wrong time and 
migrations from it give the wrong results. A brave run, perhaps, but Renfrew's 
solution is not a convincing one. 

Maintaining Renfrew's speculative vein, I would agree that there may have 
been some population movement from Anatolia to Greece and on into the 
Balkans beginning in the seventh millennium BC. I would agree that it could 
have introduced a new language, though it is always possible that pre-Neolithic 
populations on both sides of the Bosporus spoke the same or related languages. 
If they did introduce a new language, I would have expected a language similar 
perhaps to a distant ancestor of those that are historically attested in Anatolia 
such as Hattic, and that this would provide us with our earliest retrievable 
evidence of substrate languages in Anatolia and the Aegean. This language may 
well have been spread through Southeast Europe where, in the course of several 
millennia, it underwent regional differentiation in the farming villages of the 
various cultures of this region. It may possibly have been extended somewhat 
further into Danubian Europe. However, the survival of native languages after 
their gradual adoption of agriculture particularly throughout the Mediterra- 
nean, and the rest of the European periphery, would have ensured that 
Neolithic Europe was divided into numerous unrelated languages, probably 
language families. Only Etruscan, Tartessian, Iberian, Basque and Indo- 
European survived into the written record. 

If we can reject outright an association of the Proto-Indo-Europeans with the 
Early Neolithic of western Anatolia and the Aegean, we need only briefly 
comment on the Balkans as a suitable homeland. This area has been proposed 
by a number of linguists, most recently perhaps, Igor Diakonov. It should be 
noted that Diakonov, like other linguists who have postulated a Balkan 
homeland, such as Boris Gornung, exclude Greece from the homeland and 
require a later expansion from the Balkans into Greece to explain the 
appearance of the Greek language. Here I must join Renfrew in the 
conventional wisdom which makes it nearly impossible to separate the Early 
Neolithic communities of the Balkans from those of Greece since they both 
appear to have taken their origin from the same background of Neolithic 
colonization, and their differences are trivial compared with their similarities in 
both material culture and behaviour. In short, a Balkan homeland is merely an 
Anatolian-Aegean origin once removed and fails for all the same reasons. 

Only brief mention need be made regarding the archaeological plausibility of 
the Indo-European homeland recently proposed by Gamkrelidze and Ivanov 
which is set to the region of eastern Turkey, the southern Caucasus and 
northwestern Iran, roughly in the location of Armenia. First, there is really no 



archaeological evidence whatsoever for an expansion from this region, much 
less an historical pattern of migrations, that might account for the later 
positioning of the Indo-European languages. Furthermore, with the possible 
exception of Johannes Schmidt's attempt to place the Proto-Indo-Europeans 
near the Babylonians because of what he believed were similarities in their 
numerical systems, there has probably never been a solution to the homeland 
problem that has so obviously confined the Proto-Indo-Europeans in an area 
virtually surrounded, if not itself occupied, by historically attested non-Indo- 
European peoples. Gamkrelidze and I vanov have much to say that is innovative 
and interesting but their solution to the homeland problem is wholly without 
archaeological support. 

We can better examine the possibility of an archaeological solution if we 
separate the Indo-Europeans territorially into the divisions suggested earlier at 
the end of Chapter Three. First, we have the territory of Northern and 
Central Europe which, during the period that we assign to the Proto-Indo- 
Europeans, was occupied primarily by the Neolithic TRB culture and later by 
the Globular Amphora and Corded Ware cultures. The last of these occupies a 
vast territory by the third millennium which is thoroughly congruent with the 
more immediate origins of the later Celts, Germans, Baits and Slavs, and 
possibly the Italic-speaking peoples. The second major region comprises much 
of Southeastern Europe from which we would most conveniently derive the 
Illyrians, Thracians, Dacians, Greeks, and possibly the Anatolians, Phrygians 
and Armenians. It is also a likely staging area for some of the Indo-European 
peoples of Italy. Finally, we have the third major region encompassing the 
Pontic-Caspian steppe where we naturally look for the earliest Indo-Iranians 
and the most likely home of the Tocharians. We could draw a circle around this 
entire area set to 2500 bc and there would be few who would argue that we had 
not fairly encompassed all of the known Indo-European peoples or their more 
immediate lands of origin. Can we call this the solution to the homeland 
problem and leave it as settled? If we could, the next two chapters would 
certainly be far less controversial though also probably less interesting. But to 
accept such an enormous territory as the homeland requires us to spread the 
Proto-Indo-Europeans over an area of 2,000,000 square kilometres, or more. It 
would also require us to accept the inclusion within the homeland of extremely 
diverse cultures which some archaeologists insist share no historical or genetic 
connection. We really cannot avoid at least making an attempt to discover a 
smaller area that might better explain the formation of Indo-European 
languages across this larger region by about 2500 B c. Besides, there is one model 
which does claim to provide both the archaeologists and linguists with exactly 
that for which they seem to be looking. 

Although the hypothesis of an Indo-European homeland situated exclu- 
sively in the Pontic-Caspian steppe has been advanced a number of times this 
century, the present formulation of this theory owes much to the publications of 
Marija Gimbutas who has argued for over twenty-five years that the- Proto- 
Indo-Europeans should be identified with her Kurgan tradition. This term 



embraces a series of cultures occupying the steppe and forest-steppe of the 
southern Ukraine and south Russia. By the fourth millennium bc this region 
evidences all of the attributes of a putative Indo-European society recon- 
structed from linguistic evidence - including the most geographically 
indicative such as the domestic horse and wheeled vehicles. Settlements are few 
and most of the various cultures of the region are known primarily through their 
mortuary practice. This normally involves burial in an earthen or stone 
chamber, the frequent presence of ochre, and in many instances the erection of a 
low tumulus (Russian kurgan). Grave goods may include weapons and animal 
remains, especially of sheep/goat, but also of cattle and horse. The capsule 
image of the Kurgan tradition is a warlike pastoral society, highly mobile (hence 
the paucity of habitation evidence), which employed both wagons drawn by 
oxen and rode horses. These people seemed to have originated in the eastern 
steppe, perhaps in the Volga-Ural region, and to have pushed westwards until 
they burst in on the later Neolithic or Eneolithic cultures of Eastern and Central 
Europe and brought about a thorough transformation of European society. 
They also expanded southwards through the Caucasus to occupy Mesopota- 
mia; some pushed beyond towards India while others remained on the steppe 
and pressed eastwards where they passed into Central Asia, the staging area of 
,the Iranian and perhaps the Tocharian migrations. 

The expansion into Europe, the most closely argued part of the hypothesis, is 
attributed to three waves of folk movements spanning the centuries from about 
4000 to 2500 bc. The evidence for this expansion rests on a variety of novel 
traits and transformations seen in the development of European prehistory. 
Very briefly these can be summarized. 

The evidence for a change in mortuary practice plays a primary role in both 
the Kurgan theory and almost any attempt by an archaeologist to substantiate a 
claim for population movement. Into the territory of Southeastern Europe, 

85 The Kurgan homeland and its expansions according to Marija Gimbutas. 



characterized by Gimbutas as sexually egalitarian and peaceful, there appear 
alien burials morphologically identical to those on the steppe. These are 
generally confined to males and are accompanied by weapons - arrows, spears 
and knives; and by symbols of power - horse headed sceptres. The rite of suttee, 
the sacrificial execution of a woman on the death of her husband, is indicated in 
some burials suggesting the patriarchal character of the warrior-pastoralists 
who superimposed themselves on the local agricultural populations. 

The economy of the pastoralists is seen as being imported into the Danube 
region by Kurgan intruders. Gimbutas suggests that the lack of good grazing 
land in the steppe probably impelled the Kurgan migrations into Southeast 
Europe. The earlier stable agricultural economy is seen to collapse in the face of 
the intruders, and the post-Neolithic economies of the region rapidly 
emphasize stockbreeding over agriculture. 

An important component of these movements is the spread of the domestic 
horse from the steppe into the Danube region. Horse bones, cheek pieces (here 
largely limited to the northwest Pontic agriculturalists), and the horse-head 
sceptres are all cited as evidence of Kurgan intrusions. 

Earlier stable tell settlement in the east Balkans disintegrates under the 
pressure of Kurgan populations and most sites are abandoned. There is a 
general population dislocation with the final phases of a number of Balkan 
Neolithic cultures being played out in marginal lands such as islands or 
mountain cave sites. In addition, characteristic cultural markers of those 
societies, generally ceramic, appear to be pushed increasingly westwards under 
the pressure of eastern invaders. Actual settlement remains decrease with new 
more mobile communities, or else there emerge defended chieftain's 
strongholds built either in stone, such as Ezero in Bulgaria, or in wood, such as 
Vucedol in northern Yugoslavia. 

A rapid disintegration in fine ceramics, and the almost complete disappear- 
ance of painted wares, is attributed to the Kurgan people who themselves were 
content to fashion fairly crude shell-tempered cooking vessels. The copper 
industry of the Eneolithic Balkans collapses and is gradually replaced by 
arsenical-bronzes whose form and provenience are attributed to the Caucasus 
metallurgical centres via the steppe. The vector for this technological change is 
identified by long-distance Kurgan exchange. The Kurgan people also 
introduced the wheeled vehicle to the Balkans and Central Europe. 

A society identified as essentially matrifocal, that is, centred on females, and 
which is emphasized by virtually thousands of female clay figurines, abruptly 
disappears under the Kurgan warriors whose religious attention was more 
attracted to warlike sky-gods and sun worship. The new religion is especially 
seen in the erection of stone stelae in the Alpine region on which are depicted 
horses, wagons, sun bursts, and especially weapons - axes, spears, arrows and 
daggers - characteristic of the warlike society that honoured such principles. 

Finally, there is some evidence that there was an infusion of a new physical 
type into the Danube region which can easily be traced back to the steppe 
region. Here we are talking of a Proto-Europoid C type which is identical to 



skeletons examined in the steppe region and which are now found mixed with 
autochthonous Mediterranean types associated with the earlier Neolithic of 
Southeast Europe. 

These Kurgan intrusions, therefore, are credited with causing the collapse of 
the Southeast European Eneolithic cultures and their subsequent amalgama- 
tion into mixed Kurganized societies. Kurgan elements inherent in some of 
these societies then spread north and westwards where they underlie the 
transformations taking place in Northern Europe and result in the formation of 
the Corded Ware horizon. Similar evidence for migration - kurgans, horses and 
warriors - are also seen in the south Caucasus and Anatolia providing an 
intrusive culture for the Indo-Europeans of Anatolia. The expansion of such 
elements eastwards, as in the Early Bronze Age cultures of south Siberia, 
establish a Kurgan entity in the zone from which we derive Iranian peoples. 

The Kurgan solution to the Indo-European problem would thus appear to 
solve our problem economically by providing a homeland congruent with the 
Proto-Indo-European culture as reconstructed by linguistics and occupying a 
geographical situation compatible with the most plausible expansion of Indo- 
European speakers. The archaeological evidence for the expansions is not 
limited to a few traits which might be easily dismissed as the result of exchange, 
but is rather all the major features of a culture in the course of expansion into 
alien territory. The warlike society of these mobile invaders provides the 
Kurgan people with an appropriate means of expansion and an explanation for 
their success at colonizing vast areas. 

The Kurgan solution is attractive and has been accepted by many 
archaeologists and linguists, in part or total. It is the solution one encounters in 
the Encyclopaedia Britannica and the Grand Dictionnaire Encyclopedique 
Larousse. It describes Indo-European expansions in a framework congruent 
with expectation, and perhaps most importantly, it derives the Proto-Indo- 
Europeans from the Pontic-Caspian region, a territory which its bitterest 
opponents would normally admit was at least Indo-Iranian and undisturbed by 
population intrusions since the beginning of the Neolithic. Opposition to the 
Kurgan solution does not lie with those who would argue that the Pontic- 
Caspian steppe was not the territory of the earliest Indo-European speakers, 
but that the Proto-Indo-European homeland must have been larger; there is no 
alternative homeland from which archaeologists would derive all of the cultures 
of our late Indo-European territory. One might at first imagine that the 
economy of argument involved with the Kurgan solution should oblige us to 
accept it outright. But critics do exist and their objections can be summarized 
quite simply - almost all of the arguments for invasion and cultural 
transformations are far better explained without reference to Kurgan expan- 
sions, and most of the evidence so far presented is either totally contradicted by 
other evidence or is the result of gross misinterpretation of the cultural history 
of Eastern, Central and Northern Europe. In the next two chapters we will 
examine the evidence in more detail to determine whether it is in fact possible to 
isolate a more confined Indo-European homeland. 



The Archaeology of the Proto-lndo-Europeans 

Archaeology can offer nothing new to the study 
of Proto-Indo-European civilization. For any 
candidate culture advanced by archaeologists as 
the Proto-Indo-European culture, only two types 
of evidence can be offered: evidence that 
conforms to the evidence offered by linguists, 
which will be tautological (although helpful as 
support and external validation), or evidence that 
differs from the linguistic evidence, which will 
then call into question whether the candidate 
ought not to be rejected in favor of another that 
better fits the linguistic evidence. 

Bruce Lincoln 1981 

The evidence examined so far suggests that Proto-Indo-European speakers 
occupied the region between the Dnieper river on the west and the Ural on the 
east sometime during the fifth to third millennia bc. Such an assertion would 
probably be embraced by most scholars, linguists or archaeologists engaged in 
the study of the earliest Indo-Europeans. This consensus, however, swiftly 
collapses if we also assert that these two rivers formed the precise boundaries of 
the Proto-lndo-Europeans and that the homeland was effectively confined 
between 30 and 50 degrees longitude. Many claim that the border of the Proto- 
lndo-Europeans should extend west to include the territories of the TRB 
(Funnel Beaker) culture of Northern Europe and the Linear Ware culture and 
its descendants in Danubian Europe. The consequence of accepting such an 
expansion generally relegates the Pontic-Caspian region to the immediate 
ancestors of the Indo-Iranians. Other scholars would not halt the eastern 
borders of the Proto-lndo-Europeans at the Ural but extend them into 
southern Siberia. Nevertheless, whatever one's preference, there is still general 
agreement that an examination of the Dnieper-Ural region from the fifth to the 
third millennia bc should have little difficulty intersecting Proto-Indo- 
European communities. We will first review the archaeological evidence from 
this territory, therefore, and then, in the next chapter, extend our examination 
both east and west to establish the limits of the Proto-lndo-Europeans. 


Dawn of the Proto-Indo-Europeans 

By about 9000 bc the northern retreat of the glaciers across Ice Age Europe 
invited the radical alteration of the environment of the Pontic-Caspian in 
directions still recognizable today. Released from their refuge areas, a 
succession of trees colonized the lands north of the Black Sea. In general, the 
southernmost region formed a broad band of essentially steppe vegetation 
where trees were primarily confined to river valleys. To the north lay the forest- 
steppe which invited the establishment of communities of birch, pine and 
willow at first, then hazel, and finally elm and oak by the sixth millennium BC. 
These trees are, of course, all reconstructed to the Proto-Indo-European 
vocabulary although they are hardly confined to the Ukraine and south Russia. 
Nor do the animals which have been recovered from archaeological sites date 
to this period. These include dog, wild pig, wild cattle, red deer, wild horse, 
bear, fox, wolf, lynx, beaver and hare, plus an equal number of species not 
reconstructible from the Indo-European languages such as reindeer and saiga 
antelope. We may speculate that at least some of the names of these trees and 
animals were distantly ancestral to the forms we later ascribe to Proto-Indo- 
European. To argue otherwise would suggest that the Mesolithic ancestors of 
the Proto-Indo-Europeans totally abandoned their inherited vocabulary in 
later periods which is hardly probable. We may imagine, then, that during the 
Mesolithic some of the communities occupying the banks of the great river 
valleys of the southern part of the European USSR probably spoke languages 
that would later evolve into Proto-Indo-European. 

The Mesolithic period begins about 9000 BC and ends about 6000 bc by 
which time the native populations of the Pontic-Caspian began to adopt the 
manufacture of ceramics and the keeping of domestic animals. When 
examining the Mesolithic from our own particular perspective, we are tempted 
to seek some form of broad cultural uniformity that might presage the later 
formation of the Proto-Indo-Europeans. The primary archaeological evidence 
is stone tools and the determination of cultural boundaries is mainly based on 
analyzing the forms of these tools, their method of manufacture and their 
distribution as a percentage in various assemblages. For much of the Pontic- 
Caspian the stone industries are characterized by the production of geometric 
microliths or tiny flint blades. Since almost all subsequent development in this 
region appears to have been established on this Mesolithic base, some have 
perceived in this wide distribution of geometric microliths the foundation of the 
later Proto-Indo-European ethnic region, but most archaeologists today would 
be extremely sceptical of ascribing ethno-linguistic identities to vast lithic 
techno-complexes. In addition, a closer examination of this region indicates 
much greater cultural complexity. 

The best studied area of our region is the Ukraine where over 300 sites have 
been ascribed to at least twenty different Mesolithic cultures or groups. Half of 
these groups are assigned to the Early Mesolithic (ninth to seventh millennia) 
and the other ten are attributed to the Late Mesolithic (seventh to sixth 
millennia bc). These cultures can be grouped into two major cultural 



territories. The cultural groups of the southern steppe zone are technically 
similar to the broad Azilo-Tardenoisian tradition of Western Europe in their 
generally exclusive reliance on microliths. They share industrial similarities 
which link them to the Danube and the Near East as well as to their own Late 
Palaeolithic ancestors. The physical type of this southern region, attested by the 
remains in the cemetery of Voloshskoe and other sites, is described as gracile 
and Old Mediterranean. This contrasts sharply with the more robust Cro- 
Magnon physical type which is usually associated with the northern forest- 
steppe and which appears to intrude southwards into the steppe zone by the 
Late Mesolithic. The northern zone had a macro-microlithic industry which 
some would relate to the Swiderian-Kunda groups of Poland and the Baltic. 

The archaeological evidence across the Pontic-Caspian suggests that 
settlement was limited almost exclusively to the river valleys and the shores of 
what have since become dried lakes. With a hunting-gathering-fishing 
economy, populations were likely to have been both sparse and well scattered 
except where natural conditions especially favoured fishing and permitted 
larger communities. This would seem to have been the case on the lower 
Dnieper where we find nearly ninety burials in the Mesolithic cemeteries of 
Vasilevka and Voloshskoe. Here the dead were generally buried flexed on their 
sides, and ochre was occasionally employed in the burials. A number of 
individuals had suffered violent deaths, apparently from arrows. These may 
have been victims of perennial disputes over fishing rights to the most favoured 
locations. The cemeteries were almost exclusively composed of adults whose 
average age at death was about thirty-five years. 

The society indicated by the archaeological evidence for the Mesolithic 
clearly predates the type of culture that we reconstruct from the Indo- 
European languages, and we can again emphasize how nonsensical it is to chase 
Proto-Indo-Europeans back to the Mesolithic not to speak of the Palaeolithic. 
Indeed, as the physical environment of the Mesolithic was markedly different 
from that of the earlier Ice Age, it is probable that some components of the 
vocabulary concerning the environment may have only been formulated during 
the Mesolithic. We should also note that the mixture in both physical types of 
the population and cultural groups that runs throughout the Mesolithic should 
shatter simplistic delusions about any physical or ethnic purity that might be 
attributed to the Indo-Europeans. 

Emergence of Proto-Indo-European Society 

As we saw in Chapter Four, the Proto-Indo-European vocabulary clearly 
reflects an economy that emerged in Eurasia only with the Neolithic period. 
Very broadly, this form of economy which sees the exploitation of domestic 
animals and the first traces of village settlement appears in the Pontic-Caspian 
by approximately the sixth millennium bc. Most archaeologists would seek an 
external source for the earliest appearance of domestic livestock, agriculture, 
pottery and the other paraphernalia of the earliest farming communities. 

1 88 


86-88 Early Neolithic cultures in the west Pontic region. On the right are a clay pot and a 
stone vessel from the Sursko-Dnicper culture. 

The evidence for an expansion of the Neolithic economy from the southwest 
is largely centred on the origins of the Bug-Dniester culture which is known 
from about fifty sites situated primarily on the southern Bug and another ten on 
the river Dniester. By the late seventh millennium bc there begins to appear the 
first traces of domestic animals, such as pig and cattle, in what are otherwise 
typical hunting-fishing settlements. By at least the sixth millennium bc there is 
also evidence for pottery. The settlements tended to be small with nine huts on 
Baskov Island representing the most substantial site. The economy here was 
primarily based on hunting red and roe deer, wild pig, and fishing. The lithic 
technology is still firmly rooted in the preceding Mesolithic which has generally 
argued for a substantial degree of continuity in this region. The presence of 
domestic livestock on Bug-Dniester sites has normally been explained as the 
result of stimulus from neighbouring farming communities to the west. 
Similarly, the pottery has been regarded as a feature also borrowed from 
neighbouring agricultural populations. 

Apparently some slight immigration from the west and subsequent 
acculturation of the native population stimulated the creation of the Bug- 
Dniester culture. The expansion of farming populations throughout the 
Balkans had progressed to the fringes of the Pontic region where both the Cri§ 
culture in the southwest and the Linear Ware culture on the northern Dniester 
appear by the late sixth millennium. The fabric and decoration found on Bug- 
Dniester sites ties them closely to the Balkans and we may agree with Pavel 
Dolukhanov and others who characterize the Bug-Dniester culture as the result 
of hunter-gatherers who have adopted some of the forms of the Neolithic 
economy and technology from neighbours, or from small groups of farmers, 
who may have expanded out of regions of higher agricultural potential into the 
forested region of the Dniester and beyond. 

The subsequent cultural development in this western Pontic region sees 
much more intensive interactions and possible migrations from the Balkans 



that ultimately result in the emergence of the Tripolye culture. We will have 
more to say about this below. 

East of the Bug-Dniester culture this same general form of primary hunting- 
gathering-fishing economy which sees the introduction of some pottery and 
domesticated animals also emerges on the middle Dnieper. Here the poorly 
known group of Sursko-Dnieper sites are to be found. Situated on the islands of 
the Dnieper and occupying small areas which evidence semi-subterranean 
huts, the Sursko-Dnieper population also practised an economy largely 
dependent on hunting and fishing with some possible traces of domestic 
animals such as cattle. Pointed based pottery, manufactured in both clay and 
stone, is also a characteristic of this culture. The lithic industry lies in the 
preceding Mesolithic and argues again for distant contacts and the adoption of 
some Neolithic traits by otherwise predominantly hunting-gathering popula- 
tions. There are no absolute dates for this culture but it is generally correlated 
with other sites which date to the sixth millennium bc. Contemporary with 
these developments were the appearance of pottery and possibly domestic 
animals among hunting-gathering populations in the Crimea. 

It would appear that the area from the Dniester to the lower Dnieper and 
Crimea was largely one which saw the gradual adoption of Neolithic traits by 
native Mesolithic populations. The Mesolithic foundations of these early 
Neolithic cultures are largely associated with the more southerly Mesolithic 
cultures. The expansion of the Neolithic economy from the west to the Bug- 
Dniester culture can be argued on the basis of the temporal priority and 
proximity of neighbouring agricultural communities in the Balkans and 
Moldavia. However, extrapolating this process from the southern Bug to the 
Dnieper and the Crimea where there is a gap of several hundred kilometres does 
pose some problems. Why this is so critical to the problem of the Indo- 
European origins will be evident later when we have examined the origins of the 
Neolithic cultures of the Volga and Caspian area. 

By the early fifth millennium bc there emerged in the Ukraine one of the 
most extensive of the Neolithic cultures of the Pontic-Caspian region, the 
Dnieper-Donets culture. The culture is known from over 200 sites. It initially 
appeared on the middle Dnieper to the northern Donets and then expanded in 
almost all directions apparently absorbing other local Neolithic groups. The 
settlements of this culture are not well known and tend merely to repeat the 
pattern encountered elsewhere with its sparse evidence for semi-subterranean 
huts. The ceramics were initially pointed based, but flat-based wares emerge in 
the culture's later phases. Lithic technology continues the macro-microlithic 
forms of the preceding Mesolithic period although the later phases see the 
appearance of larger flint and polished stone axes and the disappearance of 
microliths. The economic evidence from the earliest stages is almost exclusively 
from hunting and fishing. The prey included wild cattle, elk, red and roe deer, 
wild pig, onager, fox, wildcat, hare and bear. In subsequent stages there appears 
to have been an increase in population and an advance southwards into the 
steppe. The later stages see the growing importance of domestic animals - 


cattle, sheep/goat, pig, horse and dog - and the initial appearance of agriculture. 

Far better known than the settlements are the approximately thirty 
cemeteries which have yielded about 800 burials. They generally exhibit a 
recurrent pattern of burial in elongated pits with groups numbering from two to 
a dozen buried together in rows, or in large collective burial pits. Individual 
burial is also known. The burial position is extended on the back with variable 
orientation, and the dead were frequently sprinkled with ochre, a custom that 
we will see continues into subsequent periods. Although ceramics are not often 
included as burial gifts, there is an assortment of other items, especially 
ornaments such as beads fashioned from shell, stone, deer or fish teeth or, very 
rarely, copper. Flint tools also occur. The most famous of the cemeteries is 
Mariupol which is located near the Sea of Azov and which yielded 122 burials. 
In addition to the grave gifts mentioned above, there were also exotic items such 
as axe-heads carved from porphyry, and copper rings, all of which attest 
exchange relationships with the Caucasus. The slight increase in grave gifts in 
the later periods has suggested to some archaeologists that we are witnessing a 
gradual growth in the complexity of social organization as agriculture and 
stockbreeding became more important in the Dnieper-Donets culture. 

The great number of burials also provides considerable evidence for the 
physical type of the Dnieper-Donets population. They are predominantly 
characterized as late Cro-Magnons with more massive and robust features than 
the gracile Mediterranean peoples of the Balkan Neolithic. With males 
averaging about 172 centimetres in height they are a fairly tall people within the 
context of Neolithic populations. 

The physical type, the extended supine burial position, the continuity with 
the preceding macro-microlithic industry, and similarities in ceramic decora- 
tion with the sub-Neolithic cultures of the northern Forest Zone have all 
suggested a northerly origin within the Ukraine, and the foremost authority on 
the culture, the Ukrainian archaeologist, Dmitry Telegin, assigns them to a 
broad cultural region that spanned the Vistula in Poland southeast to the 
Dnieper. We will later see how this might bear on establishing the limits of the 
Proto-Indo-Europeans. Conversely, we should also note that Alexander 
Formozov argues for an essentially 'southern' origin for the Dnieper-Donets 
culture that does not relate it to more northerly sub-Neolithic cultures. 

On the lower Don there also emerges a local Neolithic culture which takes its 
name from the well-stratified site of Rakushechny Yar. Several radiocarbon 
dates from this site attest the appearance of ceramics by about 5000 to 4500 bc 
among an essentially hunting-fishing population which gradually adopted 

East of the Don, the earliest Neolithic cultures are only now emerging into an 
interpretive framework among Soviet archaeologists who have been greatly 
increasing our knowledge of the eastern steppe and forest-steppe over the last 
few decades. Before this, little more was known than numerous scatters of 
pottery and flint distributed over dune sites in the middle and lower Volga 
region. These were undatable except in the vaguest sense and in the absence of 



full scale excavations of secure cultural layers they could generally be dismissed 
as camp sites of sub-Neolithic hunter-gatherers who had gained some use of 
ceramics from their western or southeastern neighbours. While this interpret- 
ation of many of the small camp sites may still stand, there are now an increasing 
amount of data that suggest that the course of the Neolithic on the easternmost 
fringe of Europe may have been far more complex and interesting than 
previously imagined. Moreover, it raises some spectacularly difficult issues 
concerning Indo-European origins. 

Stretching from the southern Urals across the Volga and on westwards to the 
Manych depression is the Seroglazovo culture of the Pre-Caspian region. It is 
attested from nearly 100 sites, mainly surface finds, which include egg-shaped 
pots and lithic industries still firmly derived from the Mesolithic. The sites are 
almost invariably situated along river or lakeside shores which would have been 
far more abundant then, when the Caspian Sea was much larger than at present. 
This provided a regimen of deltas and marshes along the southern fringe of the 
Seroglazovo culture. Radiocarbon dates for the maximum transgression of the 
Caspian suggest that the Seraglazovo culture dated to the seventh or sixth . 
millennium bc. 35 

To the north, principally in the forest-steppe of the middle Volga and 
extending eastwards to the southern Urals (the Agidel culture), there also lie a 
substantial number of recently excavated Neolithic sites. Here we speak not 
only of the recurrent remains of egg-shaped vessels or lithic traditions 
reminiscent of the Mesolithic but also of evidence for domestic livestock. 

The economy of these eastern sites was based on both hunting-fishing and 
stockbreeding. The hunted remains included the usual prey encountered across 
most of Europe and in the Proto-Indo-European vocabulary. According to 
Gerard Matyushin and Aida Petrenko, the domestic animals included the 

<Sg Early Neolithic cultures of the Caspian 
and southern Urals. 




horse, frequently the predominant species, cattle and ovicaprids. Domestic pig 
is conspicuously absent. Radiocarbon dates are not numerous and are quite 
controversial in that they suggest that the Neolithic economy had emerged here 
by at least the sixth millennium bc, if not earlier. 

In discussing the origins of this most easterly of European Neolithic cultures, 
Soviet archaeologists have often directed their attention to the southeast, 
specifically the southeast Caspian, and beyond, where one encounters the 
earliest evidence for domestic animals in the Old World. This is prompted by 
several factors. We have already seen that the earliest farming communities of 
the west Pontic do not predate the sixth millennium, and it is extremely difficult 
to tie the earliest appearance of domestic livestock on the eastern forest-steppe 
with a stimulus so distant and apparently too late. Secondly, the appearance of 
domestic sheep in the economy of these eastern sites suggests direct contacts 
with the areas of original sheep domestication in the Near East, since Soviet 
archaeologists do not regard the sheep as native to the Volga-south Ural region. 
Furthermore, the sequence of egg-shaped ceramics bears a generic similarity 
with the earliest pottery of the southeast Caspian which should date back to the 
seventh millennium BC. Hence, an explanation which sees an expansion of the 
Neolithic economy from Central Asia provides a form of solution to the 
problem of the earliest Neolithic cultures of the Volga-Ural region. It also 
offers an historical association between the easternmost Neolithic populations 
of the Pontic-Caspian and their Asiatic neighbours. Yet one can hardly disguise 
the problems inherent in such a model since the distance between the earliest 
Neolithic cave sites of the south Caspian (for example, Djebel, Dam-Dam- 
Chashma and Belt) is not only very great but also they are separated by vast 
deserts. Clear intermediaries between them and the north Caspian-southern 
Urals are not yet forthcoming although one can hardly pretend that this region 
has been extensively surveyed. The only generalization one can make about a 
south Caspian origin for the Neolithic of the Volga-Ural region is that it is hotly 
disputed. 36 

Another possible route for the Neolithic economy, especially ovicaprids, 
might lie across the Caucasus. We now have radiocarbon-dated Neolithic- 
Eneolithic settlements such as Toyretepe, Gargalartepesi, Shulaverisgora, 
Imirisgora, Kharamis Didigora and Arukhlo I that all calibrate to the sixth 
millennium bc, while the site of Shomutepe has indicated a seventh- 
millennium date. These early sites reveal faunal assemblages with ovicaprids 
totalling 50 per cent or more of the remains, followed by cattle and then pig. If 
sheep were indeed absent from the Pontic-Caspian Mesolithic, the Caucasus 
would provide a somewhat more proximate source for their introduction than 
the Balkan Neolithic cultures, especially as sheep are known from Mesolithic 
sites in Georgia. Unfortunately, evidence for early Neolithic settlement 
immediately to the north of the Caucasus has not yet been discovered. 

While the details of the initiation of the farming economy in the Pontic- 
Caspian remain a major problem, we can still make some general observations. 
The archaeological record is rather clear in reflecting quite substantial hunting- 



go Neolithic and Eneolithu sites of the Caucasus region. The Caucasus mountains are 
indicated by shading. 

gathering-fishing populations in this region, with relatively stable adjustments 
to their environment. This is particularly to be seen in the large cemeteries of 
the lower Dnieper but also elsew here in this same region. The gradual adoption 
of ceramics, then possibly of some domestic livestock together with evidence for 
cultivation, are all very much congruent with the picture of an essentially 
gradual shift in economy, the indigenous populations accepting various alien 
items into their culture as they saw the need. There is no evidence for intrusive 
farming colonists as we find in the Balkans other than that clearly marked by the 
borders of such farming cultures as the Cri§, Linear Ware and Tripolye. This is 
the eastern limit of any proposed 'wave of advance' and beyond it we have only 
indigenous communities for which there is no evidence of extermination or 
absorption by intrusive farmers. This is particularly obvious of the more 
easterly early Neolithic sites which lie up to 1,000 kilometres distant from the 
Balkan-derived cultures, and indicates that a correlation between Indo- 
Europeans and the Pontic-Caspian region appears to be one that involves 
exclusively native European populations rather than populations which took 
their origin in Western Asia at the beginning of the Neolithic. 

The possibility that the Neolithic economy and technology was introduced 
into the Pontic-Caspian region from more than one direction has certain 
linguistic implications. If one accepts a direct historical link from the Near East 
either across Central Asia or the Caucasus to the Pontic-Caspian, this might 
well provide a convenient venue for the diffusion of words associated with the 
new productive economy. But providing a solution for one desperate to account 
for a few of the apparent similarities between Proto-Indo-European and some 
Near Eastern words involving livestock and agriculture raises another problem. 



Our evidence so far appears to indicate that the Pontic-Caspian received 
Neolithic influences from the Balkans, and possibly both the Caucasus and east 
Caspian, and it is extraordinarily difficult to envisage a process by which the 
same terminology either diffused or was invented on both sides of the Pontic- 
Caspian. If we accept a dual or triple source for the Neolithic economy of the 
Pontic-Caspian then we may have to look to other processes for the apparent 
uniformity of agricultural terminology among the Indo-Europeans. A long 
process of convergence among the various Pontic-Caspian communities might 
possibly have resulted in a generally shared vocabulary among linguistically 
related groups. Alternatively, one area of the Pontic-Caspian may have 
achiev ed some form of ascendancy over the entire region thereby assimilating 
what may have been linguistically diverse neighbours. If this occurred, it 
should have taken place by the Eneolithic which marks the final period in which 
we may properly assert the existence of a Proto-Indo-European-speaking 

The Eneolithic period of the western Pontic 

The Eneolithic is generally marked by the addition of copper artefacts to the 
otherwise Neolithic technology. This distinction is fundamentally superficial 
since copper objects such as beads or aw Is also occur in cultures which have 
been universally described as Neolithic, and even during the Eneolithic period 
many cultures shown to possess copper objects did not utilize them as a 
significant part of their industrial technology. Nevertheless, the appearance of 
metals in many societies has often correlated with the emergence of widespread 
exchange relations, a more ranked social system and, at least in a European 
context, also a rise in warfare, defensive architecture, increased dependence on 
the pastoral component of the economy, horse domestication and wheeled 
vehicles. Both the lexico-cultural evidence of the Indo-European languages and 

gi Eneolithic cultures of the 
west Pontic region. 



our best estimates for the terminal period of Proto-Indo-European existence 
direct our attention to the Eneolithic. 

During this period the western frontier of the Pontic-Caspian w as delimited 
by several different cultures. In the far southwest, towards the Danube, was the 
Gumelnitsa culture which occupied eastern Romania and Bulgaria and whose 
continuity with the preceding cultural evolution of the Balkan Neolithic lies 
essentially unassailable. It is for precisely this reason that we have excluded it as 
non-Indo-European since it would require us to ascribe to the entire Balkans an 
original Indo-European identity which we have found incompatible on other 

The major western neighbour of the Pontic-Caspian Eneolithic peoples was 
the Tripolye culture, the w estern variant of the Cucuteni-Tripolye continuum 
that spanned Romania to the Ukraine and whose existence ranged over 1,500 
years. In the Ukraine alone there are approximately 1,000 sites of this culture 
known. These range from small settlements comprising only about a dozen 
houses to vast settlements whose structures number in the hundreds. The 
increasingly defensive nature of these settlements throughout the course of the 
Eneolithic, and the intensified contacts with their eastern neighbours, will 
demand comment in the next chapter. At present, our interest is primarily 
drawn to the origins of the Cucuteni-Tripolye culture. 

Although some have claimed a local derivation rooted in the earlier Bug- 
Dniester culture, both the accumulation of evidence over the past decades and 
general opinion now emphasize the essentially Balkan roots of this culture. It 
first appears in Romania where the Pre-Cucuteni I phase dates to about 5000 
bc. The earliest sites find their closest parallels in the architecture, ceramics, 
figurines, lithics and economy of the Boian culture of Romania. It is from this 
region that the Cucuteni-Tripolye culture seems to have expanded eastwards 
into the Prut-southern Bug region where Tripolye sites begin to emerge about 

92 Reconstruction of a Tripolye settlement. 



4500 bc; it then spread both northeastwards towards the Dnieper and 
southwards towards the northwest shores of the Black Sea. The initial 
extension into the Ukraine would naturally have carried these people into the 
earlier territory of the Bug-Dniester culture, but the marked distinctions 
between the two have generally indicated that the agriculturally more advanced 
Tripolyeans virtually assimilated the preceding inhabitants. Therefore, the 
Cucuteni-Tripolye culture, like the Gumelnitsa culture to its south, appears to 
derive from the great Balkan block of Neolithic cultures, which we have good 
reason to set outside the range of Proto-Indo-European candidates. 

North of the Tripolye culture, on the northwest border of the Ukraine and 
Moldavia, the Eneolithic is represented by the Lengyel and TRB cultures. As 
both of these occupied very substantial regions in which the putatively Indo- 
European Corded Ware horizon subsequently appears, we have no a priori 
grounds to exclude them from the Proto-Indo-European world. We will 
examine their relationship with the Pontic-Caspian region in the next chapter. 
It is now time to examine those Eneolithic cultures whose origins would appear 
to lie with the indigenous populations of the Pontic-Caspian and who have the 
surest claim to being designated Indo-Europeans. 

The Early Eneolithic of the Pontic steppe and forest-steppe 

The Eneolithic cultures of the Pontic-Caspian emerge by 4500 BC and 
subsequently evolve into Early Bronze Age cultures by about 2500 bc, if not 
earlier. Since these dates coincide almost precisely with the suspected floruit of 
the Proto-Indo-Europeans before the emergence of distinct Indo-European 
groups, we are justified in examining this period more closely than the 
preceding one. Our attention will focus on those areas of material culture or 
behaviour that seem to pertain most to our reconstructions of Indo-European 

In broad terms, all of the cultures which we are about to encounter take their 
origin from their Neolithic antecedents in the Pontic-Caspian although precise 
derivations are very much open to debate. The earlier Eneolithic cultures 
embrace a series of individual cultures, all of which are distinguished in 
archaeological nomenclature, but which also display recurrent traits that point 
either to long-standing mutual contacts or underlying genetic relations. These 
include the Sredny Stog, Novodanilovka, Lower Mikhaylovka-Kemi Oba 
cultures in the west and the Samara, Khvalynsk and southern Ural Eneolithic 
cultures in the east. In addition, residual Neolithic cultures such as the 
Dnieper-Donets culture lingered in some regions to be contemporary with 
some of the Early Eneolithic cultures before finally being absorbed by them. By 
the end of the Eneolithic period, about 3500-2500 bc, most of the Pontic- 
Caspian region was occupied by the Yamnaya (Pit-grave) horizon. The 
successors of the Yamnaya culture are set to a period in which one generally 
assumes the emergence of already differentiated Indo-European-speaking 
peoples. It is during the course of the Eneolithic period that those who support 



gj Distribution of the Sredny Stog and Novodanilovka cultures. 

the Pontic-Caspian as the exclusive homeland of the Indo-Europeans argue for 
their initial expansion. 

The Sredny Stog culture is easily the best known of the Early Eneolithic 
cultures of the Pontic-Caspian. It is attested by approximately 100 sites located 
primarily along the middle Dnieper and extending eastwards to include the 
Donets and the lowermost reaches of the Don. Both radiocarbon dates and 
synchronisms w ith the better-dated Tripolye culture indicate that this culture 
flourished about 4500-3500 bc. 

The evidence for settlements in the Sredny Stog culture is not extensive and 
the site of Dereivka on the middle Dnieper is by far the most impressive. An 
area of over 2,000 square metres was apparently bordered by some form offence 
which enclosed several houses, work places and areas of ritual activity. The 
houses were slightly sunk into the ground, rectangular in shape, with the largest 
measuring 13 by 6 metres. Hearths were found within them. Scattered about 
the site were various activity areas including a place for repairing fish gear and 
processing one's catch, a potter's workshop, and a place where bones were 
worked into tools. Semi-subterranean huts are also known from Alexandria, 
and surface dwellings, similar to those from Dereivka, have been uncovered at 
Konstantinovka on the lower Don. At Dereivka, in addition to more secular 
activities, there was also some evidence for ritual. This included a deposit 
consisting of the head of a stallion accompanied by the left footbones of a horse, 
the remains of two dogs, and a figurine in the shape of a wild boar. Beneath 
one of the walls of a house was discovered a pit with the burial of a dog. 

The economy of the Sredny Stog culture was based on stockbreeding, 
agriculture, hunting and fishing. The domestic livestock, in terms of the 



()4 Plan of the Sredny Stog site of Dereivka. The area indicated with solid lines was heavily 
disturbed without proper excavation. 

attested number of individuals from five Sredny Stog sites, included horse, 
sheep/goat, cattle, pig and dog. The horse served both as a meat source, 
evidenced by the slaughter patterns which specifically reflect the butchering of 
young males, and for transport. Antler cheekpieces for fixing the bit in the 
horse's mouth are known from Dereivka and other Sredny Stog sites. This 
evidence, coupled with the logical requirements of controlling herds of horses 
from horseback, supports the thesis that the horse was probably ridden at this 
time. Generally, archaeologists are sceptical that the horse could also have been 
employed for traction. Both the small size of the animals (average withers 
height of 136 centimetres) and the absence of a suitable harness would have 
rendered these earliest domestic horses poorly suited to pull the heavy timber 
carts or wagons that are first attested in the Eneolithic. 

Although the horse would have revolutionized transportation and the 
mobility of the Sredny Stog people, it is questionable w hether it alone provided 
the complete prerequisites for true steppe nomadism involving the cyclic 
exploitation of steppe pastures on a year-round basis. While sheep were well 
suited to the steppe, the presence of domestic pigs on Sredny Stog sites points to 
a settled regime, although one that was amassing the technological and 
economic requirements which might permit specialized nomadic pastoralism. 
It is always very difficult to evaluate the comparative values of stockbreeding 
versus agriculture in prehistoric sites, yet the ev idence for half a dozen querns 
and about a dozen grinders from Dereivka at least indicates the processing of 
plant foods, although these may not necessarily have been domestic. In general, 
Soviet archaeologists assume that stockbreeding provided the primary core of 
the Sredny Stog economy. 



The wild animals hunted by the Sredny Stog population included a wide 
range of prey attested throughout Europe. The primary game animals were red 
and roe deer, elk, wild boar and, again reminding us more forcefully of the 
riverine environment of these sites, beaver and otter. Badger, wolf, fox and hare 
are also represented in the faunal samples. Birds such as mallard, pintail duck, 
goose, teal and coot were recovered as well. 

Finally, there can be little doubt that fishing also played a role in the Sredny 
Stog economy. The discovery of net sinkers, fishhooks, and fish remains at 
Dereivka confirms the exploitation of wels, perch, roach, red-eye, carp and 
pike. Unio and Palludino shells were also collected extensively. 

The technology of the Sredny Stog culture provides several areas for 
comment. Ceramics were still bag-shaped with rounded or pointed bases which 
clearly reflect their local forest-steppe ancestry. The fabric of the vessels was 
frequently tempered with crushed shell which provides a useful technological, 
and some would claim ethnic, marker for the Pontic-Caspian region. By the 
later period of the Sredny Stog culture, the so-called Dereivka period following 
4000 bc, cord ornament appears on the pottery setting a pattern which 
subsequently emerges over much of the rest of Europe at a later date. The 
implications of this design technique will assume greater importance when we 
attempt to trace the expansion of the Indo-Europeans. 

The lithic remains of the Sredny Stog culture contain much of what we 
might expect of most Eneolithic societies across Europe. Knives, scrapers, 
arrowheads, spearheads are all known. Among the antler tools, the most 
interesting are the hammers and mattocks which occur in great abundance at 
Dereivka. The excavator, Dmitry Telegin, argues that the hammers were close 
range weapons and served an analogous function to the stone battle-axes which 
appear later in the Eneolithic. Cheekpieces for horses were also fashioned from 

()5, 96 Antler cheek pieces and wild 
boar figurine from Dereivka . 



Copper objects are rare in the Sredny Stog culture and generally consist of 
little more than beads, although Telegin suggests that copper tools did exist at 
Dereivka as evidenced by traces of copper oxides on worked bone and the 
requisite techniques employed in working antler into tools. Evgeny Chernykh 
has demonstrated by spectral analysis that the copper was derived from the 
Balkan-Danubian region and probably passed eastwards via the Tripolye 
culture to the Sredny Stog. 

A number of Sredny Stog cemeteries have been uncovered and reveal 
relatively uniform burial rites. The graves are simple pits without any evident 
surface marker such as a mound. The deceased were buried on their backs but 
with their legs flexed (in distinction to the customary extended positions of the 
Dnieper-Donets culture), and they were frequently strewn with ochre. Grave 
gifts were few and included pots and tools. At Dereivka one of the graves was 
accompanied by a pot imported from the Tripolye culture. 

The disposition of the burials within the cemetery is also distinguished from 
the previous Dnieper-Donets culture where numerous burials were interred in 
a series in group pits. In the Sredny Stog culture the burials tend to be grouped 
into small numbers, two to five, which are separated from other small groups 
within the same cemetery. Soviet archaeologists see in this a shift in social 
organization where the various small groups indicate family or other kinship 
units who retained their distinct identities even within their cemeteries. 

The Sredny Stog cemeteries also provide information about both the 
physical appearance and life-spans of the population. In general, the Sredny 
Stog people are described as proto-Europoids of medium to tall stature, more 
gracile than the Dnieper-Donets people but still quite robust when compared 
with their contemporaries in the Tripolye culture. From the small Sredny Stog 
cemetery on Igren island, Ina Potekhina has examined the demographic 
structure of the population. Males who had achieved adulthood died on average 

97, g8 Pointed-based pot and antler 
'hammer-axes from Dereivka. 



about the age of thirty-six years while the few females in the cemetery, contrary 
to the usual pattern observed in prehistoric cemeteries, outlived the men and 
died at about forty-four years of age. If the very high rate of infant mortality is 
taken into consideration, the average age at death for the entire population was 
about twenty-seven years. 

There are quite marked similarities between the burials of the Sredny Stog 
culture and those of the Novodanilovka group which also occupied the lower 
Dnieper and the steppe region of the Ukraine contemporary with the early 
Sredny Stog culture. The Novodanilovka burials are grouped into small 
cemeteries, generally not exceeding half a dozen burials, among which Chapli, 
Yama, Voroshilovgrad and Petro Svistunovo are the best known. The burials 
are in the supine position with legs flexed, orientation is to the east or northeast, 
and the deceased was sprinkled with ochre - all of which are characteristics also 
seen in the Sredny Stog culture. Where they differ is in the elaboration seen 
both in the construction of the tombs and in the grave goods. Generally, the 
grave pits are lined with stone slabs and the burials are richly accompanied with 
a wide assortment of goods. Included among these are stone tools of which long 
flint knives, arrowheads and spearheads are quite typical. The flint is of high 
quality and was obtained from the Donets region. Polished stone axes, 
fashioned from slate or serpentine, maces made from stone, antler or even 
copper also occur. About a dozen copper bracelets are known; their function is 
indisputable, as they have been found about the lower arm bones of a burial at 
Chapli. Other ornaments include rings, beads and pendants of copper, and 
pendants fashioned from boar's tusk, animal teeth and shell. Globular vessels 
with slightly pointed bases also occur with the burials. 

The interpretation of the Novodanilovka group is made extremely difficult 
because of the total absence of any settlement sites. Other than a few hoards of 
copper or flint objects, such as the Goncharovka hoard with 150 knife-like 
blades, the Novodanilovka burials are without a clear cultural context. They 
were initially assigned to the Sredny Stog culture but are now recognized by 



Ukrainian archaeologists as an independent group, putatively composed of 
specialist flintworkers engaged in the long-distance exchange of flint and 
possibly copper. That future discoveries within the Sredny Stog settlements 
may extend the range of its objects and practices to include the more exotic 
items found in Novodanilovka burials is always possible, and it has been 
suggested that these graves merely represent w ealthier members of the Sredny 
Stog culture. What is most important is the pattern of broad general similarity 
between the different cultures and groups occupying the Dnieper-Ural area. 

The third major Ukrainian cultural entity of the earlier Eneolithic is the 
Lower Mikhaylovka-Kemi Oba culture which spans the region between the 
lower Dnieper and the Crimea. The lower Dnieper variant, the Lower 
Mikhaylovka culture, synchronizes roughly w ith the later part of the Sredny 
Stog culture, w hile the Kemi Oba culture of the Crimea extends into the later 

Some settlements have been ascribed to the Lower Mikhaylovka culture, the 
most famous being the eponymous site of Mikhaylovka where the Lower 
Mikhaylovka remains underlie the later Yamnaya settlement and fortifications. 
Here semi-subterranean houses were found, one of w hich measured about 1 5.5 
by 5 metres. The faunal remains from Lower Mikhaylovka were not abundant 
but include sheep/goat, cattle, horse, pig and dog, in that order, with some 
traces of hunting. The ceramics are different from the Sredny Stog culture and 
are flat-based with raised necks. Of special note is a low pedestalled bow 1 w hich 
may be interpreted as a censer, analogous to similar ritual paraphernalia 
encountered regularly in this region throughout later prehistory. 

Burials and associated rituals have attracted special attention. The burials are 
placed in low mounds (kurgans) and the presence of stone rings, cromlechs, is 
frequently noted. Hearths have been discovered built on top of the kurgans, on 
their periphery or within the burial pit itself. Grave goods are rare but may 
include pottery, copper awls or shell ornaments. 



ioj-106 The distribution of the 
Lower Mikhaylovka-Kemi Oba 
cultures. Above is a burial from 
Kemi-Oba. Left are a Lower 
Mikhaylovka pot and censer. 

One of the more striking recent discoveries of the Lower Mikhaylovka group 
is the existence of altars or offering places. Beneath a kurgan at Kalanchak was 
found a circular area on which lay the fractured remains of an anthropomorphic 
stone stela with traces of ochre; potsherds; and animal bones. Similar deposits 
have been found elsewhere. 

To the south, in the Crimea, are the remains of the Kemi Oba culture which 
is primarily represented by small cemeteries. Besides those features which are 
similar to the Lower Mikhaylovka group, for example, kurgans, cromlechs, 
eastern orientation, and so forth, there are several other features of considerable 
interest. A number of the tombs which have been built as stone cists have 
included painted ornament on the walls. Of greater representational interest are 
the carved stone stelae on w hich are depicted the heads and arms of figures, and 
which are covered with both geometric and more realistic ornament. A fine 
example of this is the stone stela that derives from Kernosovka. The stela stood 
1.2 metres high and depicts the head, including a face with a moustache and 
beard; arms; and phallus. On the front surface of the stela are carved images of 
what have been interpreted as tools such as mattocks, a battle-axe, and animals 
including two horses. There are about seventy such figures known from the 
Pontic region. Considerable evidence exists that they were employed in Later 
Eneolithic burials, especially in the construction of Yamnaya graves where they 



were used to cover the deceased. This was clearly not their original purpose 
since they were constructed to stand upright, and Dmitry Telegin suggests that 
they were originally manufactured by the Lower Mikhaylovka-Kemi Oba 
culture and later appropriated by Yamnaya tribes who reused them in their own 

These major cultural groups - Sredny Stog, Novodanilovka and the Lower 
Mikhaylovka-Kemi Oba cultures - all constitute the primary Eneolithic 
cultures of the lower and middle Dnieper to the lower Don region in the period 
4500—3500 bc. Their origins are by no means clearly understood though no one 
would deny a strong degree of continuity from the preceding Surski and 
Dnieper-Donets cultures in their development. But other impulses, seen in 
ceramics, metal working, and the elaboration of burial ritual by using stone, 
point to more distant contacts with the Tripolye culture to the west, and 
especially the Eneolithic cultures of the north Caucasus to the southeast. 

We have already noted the potential importance of the Caucasus in 
stimulating the Neolithic economy in the Pontic-Caspian region. The 
association of the northern Caucasus with the Pontic-Caspian is much more 
clearly seen in the Eneolithic period. One of the earliest north Caucasian sites of 
importance is the cemetery at Nalchik. Here were found 147 burials placed 
under very low mounds which together formed an extensive low kurgan 
covering an area of about 300 square metres. Although twelve of the burials 
were found in the supine position with legs flexed (as we frequently encounter 
in the steppe), the majority were deposited on their sides, males on their right 
and females on their left. Ochre frequently accompanied the burials. Grave 
goods included pendants fashioned from animal teeth, flint tools, and a series of 
marble bracelets. The earliest burials at Nalchik are dated to the Eneolithic. 
Other than a few other burials and a single settlement site, there is little local 
context for the Nalchik cemetery which appears to straddle the world of both 
the steppe and the Caucasus. 

Nalchick precedes the Maykop culture which takes its name from the famous 
royal barrow at Maykop southeast of the Sea of Azov. The massive quantity of 
gold and silver ornaments and vessels has long been the subject of 
archaeological debate: what were their precise chronological and cultural 
relations with the Bronze Age cultures of the Near East, Anatolia, and their 
neighbours in the Caucasus? The sites of the Maykop culture appear to cluster 
in the Kuban region from whence they extend eastwards across the northern 
Caucasus. Burials are typically found beneath kurgans which generally employ 
stone constructions such as cromlechs and stone cists. The deceased are found 
buried either in the supine position with legs flexed, or on their sides. Copper 
objects are a frequent burial accompaniment. 

The origins, interpretation and absolute dates of Nalchik and the Maykop 
culture are perennial topics of debate. Their origin, for example, is variously 
attributed to a yet unidentified local Neolithic population, or to a northward 
expansion of the Eneolithic cultures of the Caucasus. More important from our 
point of view is the elaboration of their burials, with stone constructions which 



many archaeologists see as a source for the stone-built tombs encountered in the 
Lower Mikhaylovka and Kemi Oba cultures. Similarly, they offer ceramic 
parallels for some of the cultures on the steppe. Finally, in their strategic 
position between the steppe and the major metallurgical centres of the 
Caucasus, the northern Caucasus becomes an important factor in the cultural 
development of the Pontic-Caspian in the Later Eneolithic and Bronze Ages. 

Early Eneolithic in the East 

The Eneolithic successors of the earlier Seroglazovo culture are the Samara 
culture of the middle Volga forest-steppe, and the Pre-Caspian culture to its 
south. Both cultures are still very poorly known and their formulation as 
cultural entities is relatively recent. The Samara culture, which takes its name 
from the river Samara, was only discovered in 1973. Its major site is the 
cemetery of Sezzhee where many of the practices and some of the grave goods 
encountered in the Dnieper-Donets culture are paralleled. The burials are in 
flat graves, extended on their backs, and often powdered with ochre, especially 
the graves of children. The majority of graves were accompanied by goods that 
included polished stone axes, shell beads, pendants of animal teeth, bone tools, 
ceramics, and small plates fashioned from boar tusk or shell which would be 
sewn on garments, a practice most notably attested in the Dnieper-Donets 
cemetery at Mariupol. In addition, figures carved from tusk or bone also occur 
in the graves. They were fashioned into the shapes of horses, cattle and ducks. 
The possibility that the horse was employed ritually in the burial rite is 
suggested by the discovery of horse skulls and other bones in the overburden of 
the cemetery. 

The distinctive shell-tempered Samara ceramics are known on other sites 
throughout this region and, according to Igor Vasiliev, ceramically influenced 



the forest cultures to the north. This provides a point of mutual contact between 
a segment of what we presume to have been Proto-Indo-European speakers and 
the region most often favoured as the probable homeland of the Uralic 

The Pre-Caspian culture to the south is very poorly known with little more 
than twenty sites identified. These are generally on heavily eroded dune 
surfaces where material from different periods has been mixed together. The 
sites are, as a rule, situated along the shores of dry lakes and are composed of 
flat-based ceramics, quartzite tools and occasionally animal bones. Both the 
Samara and Pre-Caspian cultures are synchronized with the Dnieper-Donets 
culture to the west. 

The successor to both the Samara and Pre-Caspian cultures is known as the 
Khvalynsk culture which takes its name from the major cemetery of Khvalynsk, 
situated on the right bank of the Volga. This cemetery covered an area of about 
1,100 square metres and revealed the remains of 158 burials. The cemetery 
reflects striking similarities with both the Dnieper-Donets and Sredny Stog 
cemeteries. While there are forty-five individual burials, the majority were 
placed in group pits ranging from a pair to as many as seven together. The 
burials were normally in the supine position w ith legs flexed, often covered with 
ochre, and orientated from north to east. The graves were simple pits, though a 
number had been covered with stones. 

The grave goods from the Khvalynsk cemetery are exceedingly rich and 
include about fifty pots, again employing crushed shell temper; beads fashioned 
from Unio shell, bone and stone; dentalium shell pendants; stone arrow heads 
and axes; bone harpoons, fishhooks and knives; and animal bones. Here, too, 
were uncovered figures carved from boar tusk and shell, and about forty copper 
objects. These included spiral bracelets and rings and, like the Sredny Stog 
copper, spectral analysis indicates that the copper was originally derived from 
the Balkan-Danubian region far to the west. In the overburden of the graves 
there were found the bones of domestic horse, cattle and sheep/goat. 

iog Distribution of Eneolithu 
cultures of the Caspian-middle Volga 

108 Horse figure from Sezzhee 
(L. c. 12 cm). 





Iff \ f 


C3 c=3> ^? 

//o Comparison of material from the Samara and Dnieper-Donets cultures. 

in (Opposite,/ material from the Khvalynsk and Sredny Stog-Novodanilovka cultures. 





O Qi Of Q 

g ©00 O 7 : 5? ',| 

gj • o> Oil 




Igor Vasiliev points out that, excluding the differences in ceramics, there are 
striking similarities in burial ritual and technology between the Khvalynsk and 
the Sredny Stog cultures. To these we may add, naturally, the presence of 
domestic horse, which is apparently known as early in the middle Volga-south 
Urals as on the middle Dnieper. The similarities suggest to Vasiliev that there 
was a broad Sredny Stog-Khvalynsk horizon embracing the entire Pontic- 
Caspian during the Eneolithic. This, he suggests, replaced an earlier broadly 
uniform horizon, associated with the Dnieper-Donets culture, that not only 
occupied the west but also evidenced strong influences on the middle Volga 
Seroglazovo and Samara cultures. 

In attempting to explain why we should have such widely similar material 
and ritual behaviour across the entire Pontic-Caspian, Vasiliev touches on the 
issue of cultural priority. In the Ukraine, the later phases of the Dnieper- 
Donets culture appear to coexist with the Sredny Stog culture until they are 
finally absorbed. In such a situation, according to Vasiliev, it is doubtful that 
one can argue for an entirely local evolution from the Dnieper-Donets to the 
Sredny Stog. Rather, he directs our attention eastwards to the middle Volga 
where the transition between the intervening cultures of the Neolithic and 
Eneolithic clearly indicate local development. The possibility that the Sredny 
Stog-Khvalynsk horizon was achieved by impulses moving from east to west 
can be proposed even if Ukrainian archaeologists emphasize what they perceive 
to be continuity between the Neolithic and Eneolithic. Indeed, this hypothesis, 
that the Ukrainian Eneolithic cultures were in part derived from movements 
from the Volga, has been argued by a number of Soviet archaeologists, though 
previously there was little chronological control of the data, nor was there the 
evidence for cultural development on the middle Volga. 

We are still far from understanding precisely why there should have 
developed such a broad band of similar cultures across the Pontic-Caspian 
during the Early Eneolithic. Obviously, a general Drang nach Westen would 
help explain the uniformity of the stockbreeding vocabulary of the Proto-Indo- 
Europeans, but this should not be accomplished at the expense of other 
pertinent archaeological data. It is clear that metallurgy was diffusing in the 
opposite direction, and we must envisage a very broad area which formed a 
sphere of constant mutual relations, especially with regard to exchange. The 
increased mobility produced by the domestication of the horse was also 
probably an important factor. The extent to which actual folk movements were 
involved is not yet clear, but they can hardly be dismissed. The ultimate result 
of these interactions and possible movements is to be seen in the Late 
Eneolithic. Here the greatest similarities arise with the Yamnaya culture, one of 
the major cultural-historical entities of prehistoric Europe. 

The Yamnaya cultural-historical area 

The final Eneolithic culture of the Pontic-Caspian region, and the last cultural 
entity which may putatively be assigned a Proto-Indo-European date, is the 



Yamnaya (Pit-grave) culture. The major floruit of this culture, substantiated by 
more than seventy radiocarbon dates, is about 3600-2200 BC. Its territory 
embraced the entire Pontic-Caspian from the Bug and Dniester rivers on the 
west across to the Ural and Emba rivers on the east. Such a territory, stretching 
3,000 kilometres across, is so vast that many archaeologists accept the 
terminology of Nikolai Merpert and refer to a Yamnaya cultural-historical area 
rather than to a single culture. 

The immediate origins of the Yamnaya culture are complicated and still very 
much disputed, although there is general agreement that both the Sredny Stog 
and Khvalynsk cultures were the primary foundations for the Yamnaya groups 
of their respective regions. While these certainly underlie some of the local 
Yamnaya variants, it should be noted, however, that Merpert envisages nine 
different regional variants in this vast continuum. 

The Yamnaya culture is overwhelmingly evidenced by the remains of its 
burials rather than from settlements. Evidence for occupation sites is almost 
unknown in some of its regional variants, and where they do occur in sizeable 
numbers, they frequently tend to be insubstantial camp sites suggesting a 
mobile form of economy. This is especially true in the steppe region of the 
Volga where Yamnaya remains are recovered from the same type of dune camps 
as in the preceding Neolithic period. 

There are, however, some major exceptions to this pattern, especially along 
the lower Dnieper where the site of Mikhaylovka offers the most substantial 
remains of a Yamnaya settlement. Here, overlying the earlier Lower 
Mikhaylovka phase, were stratified two substantial phases of the Yamnaya 
culture. The earlier occupied an area of about 1,500 square metres. Both semi- 
subterranean and surface structures, the latter with stone foundations, were 
recovered, along w ith a great quantity of ceramics, tools and faunal remains. 
The later Yamnaya phase saw the expansion of the settlement to cover an area of 

112 The regional groups of the Yamnaya cultural-historical region. 



1.5 hectares, and the erection of fortifications consisting of both a ditch and 
stone walls which still stand to a height of 2.5 metres. Several varieties of solidly 
constructed houses were discovered, including both small oval-shaped 
dwellings similar to the previous phase, and large rectangular houses with one 
to three rooms. These structures were built on stone foundations up to a metre 
high and then completed in wood and daub. Although Mikhaylovka is the 
largest of the fortified Yamnaya sites it is not unique, and several other stone 
fortified sites are known. 

Mikhaylovka offers by far the largest single sample of economic remains 
from the Yamnaya culture. Here were recovered over 50,000 identifiable bones 
of domestic and wild animals. The domestic fauna consisted primarily of cattle 
and then sheep/goat, with substantial remains of horses and some pig. The most 
typical wild fauna included the onager, red deer, aurochs, wild boar, saiga (the 
steppe antelope), and a variety of other species such as otter, fox, wolf, hare and 
beaver. Querns and flint sickle blades indicated that agriculture was practised. 
The faunal remains from the site of Repin on the Don were primarily those 
of domestic horse. 

While the existence of agriculture in the Yamnaya culture is not disputed, the 
general opinion is that the culture was overwhelmingly centred on 
stockbreeding which may have become so specialized in some of its regional 
variants that it permitted pastoral nomadism. Valentin Shilov has called 
attention to the natural conditions of the open steppe where salty soils and sands 
w ould have precluded any serious development of agriculture yet would have 
provided excellent conditions for pastoral economies. I* is in precisely these 
areas that Yamnaya camp sites have been encounters u the Volga-Ural 
region. Furthermore, Yamnaya burials are known from the deep steppe far 



from the major rivers that might have provided the regime necessary for stable 
farming settlements. Although the primary source of faunal remains for much 
of the Yamnaya region is to be found in grave offerings, these do indicate a 
marked predominance of ovicaprids, precisely the animals that would best 
exploit the steppe environment. Even in areas such as the lower Dnieper, 
Ukrainian archaeologists argue that such sites as Mikhaylovka may have served 
as centres on which camp sites of semi-nomadic pastoralists depended. 

The general picture of the Yamnaya economy is varied and dependent on the 
natural conditions in which its populations found themselves. In the major 
river valleys, where agricultural soils and forested environment provided the 
necessary basis for mixed farming settlements, the Yamnaya culture appears to 
have followed such an economy. Nevertheless, the increased development of 
stockbreeding, especially the utilization of both the sheep and domestic horse, 
assisted in the expansion of human settlement out from the river valleys into the 
deep steppe. Another obvious factor in the development of mobile economies 
was the invention of wheeled vehicles. 

Wagons are first attested in the Pontic-Caspian during the Yamnaya period. 
Quite numerous remains of wheels, and even some entire wagons, have been 
recovered from Yamnaya burials. They show the use of both the two-wheeled 
cart and the four-wheeled wagon. There is consensus that the means of traction 
was oxen, as the vehicles with their heavy solid wooden wheels would have been 
far too heavy for horses to pull at this time. The wagon is traditionally seen as 
one of the prerequisites for successful exploitation of the open steppe, since it 
provided the necessary mode of transport for both family and property which 
was necessary in a mobile economy. Horse riding, which we have already seen in 
the Sredny Stog culture, is also evidenced in the Yamnaya culture. A pair of 
wooden cheekpieces was recovered from a Yamnaya kurgan at Vinogradovka, 
near Odessa. 

The range of Yamnaya technology is not extensive. In addition to the variety 
of flint tools employed in the subsistence economy there is also a range of 
weapons - for example, flint arrowheads, daggers, stone battle-axes and maces. 
Bone and antler tools such as mattocks, harpoons and awls have also been 
recovered. Ceramics vary according to region but in some instances continue 
the pattern of shell and sand-tempered wares with decoration executed by cords 
and comb stamping. There is more copper than in earlier periods, with the 
production of awls, knives, chisels and adzes. While previous metallurgical 
developments appeared to be the result of long-distance contacts with the 
Balkan-Danubian region, we now find the beginnings of localized Yamnaya 
metallurgical centres. In the lower Dnieper copper appears to have been 
acquired from the Caucasus region and developed under its stylistic influence, 
although with its own independent production, as evidence for copper working 
at Mikhaylovka suggests. In the Volga-Ural region local copper resources 
began to be exploited. 

The primary evidence for the Yamnaya culture, as mentioned above, is 
burials. In general the burial ritual involved the digging of a shaft (the yama 'pit' 



grave) and depositing the body of the deceased, at least in the earlier periods, on 
the back with legs flexed and head oriented east or northeast. Some extended 
supine burials are also known. The deceased might lie directly on the floor of 
the pit, but there are also frequent traces of wooden planks or reeds and rushes 
being employed as a floor. The use of ochre is quite frequent and hence the 
culture is often termed the Ockergrabkultur (ochre-grave-culture) in German 
archaeological literature. The burial might be surrounded by a cromlech, or 
covered with stones, or timber planks might form a roof over the pit and a 
hearth might be placed next to the burial. The most notable feature, however, 
was the erection of a kurgan over the grave. Into this kurgan might be deposited 
subsequent Yamnay a burials or indeed burials from later periods extending all 
the way to the Middle Ages. In terms of grave goods, these might range from 
none, especially with the majority of typologically early graves, to an impressive 
quantity of items. Goods might include pots, copper knives and awls, boar-tusk 
pendants, and an assortment of bone and stone tools such as flint sickle blades, 
scrapers, stone axes and harpoons. Occasionally bird bones, generally 
interpreted as primitive flutes, have been found. Wheeled vehicles or individual 
wheels might accompany a burial. 

Animal bones are an intriguing accompaniment to many burials and the 
principal species represented were ovicaprids, cattle, horse, dog and some wild 
animals. These remains may often be interpreted simply as joints of meat 
presented as food offerings; however, other rituals were also at play. Frequently 
the skull and forelegs of a sheep, or much more rarely of a horse, are 
encountered in a grave and indicate the presence of a 'head and hooves' cult. In 
some cases the forepart of the animal might have been erected directly over the 



burial. Knucklebones of sheep were also found frequently. Knucklebones are, 
of course, a familiar gaming device, and the association between the 
knucklebone, or astragalus, and words for dice is known in various Indo- 
European languages. Their presence in Yamnaya burials may be explained as 
offerings of gaming pieces but one should also note that they show a very strong 
correlation with the burials of young children. At Berezhnovka on the Volga, 
for example, one child was accompanied by sixteen astragali which may have 
served as toys or ornaments. 

The origin of the Yamnaya culture is still a topic of debate. Essentially, the 
major issues concern which area of the Pontic-Caspian exhibits the greatest 
continuity of culture between the earlier and later Eneolithic, and whether 
chronologically earlier burials might be attributed to a particular region. Some 
archaeologists, such as Igor Vasiliev and Marija Gimbutas, argue that the 
earliest Yamnaya burials occur in the Volga region, and that the evidence for 
continuity between the Neolithic across the Eneolithic is so unassailable that 
one must attribute a priority to this region with regard to Yamnaya origins. 
Others, such as Dmitry Telegin, would find that the Yamnaya burials of the 
Ukraine are so distinctly related to the preceding Sredny Stog culture that it is 
unnecessary to seek an external origin. Although there are a great number of 
radiocarbon dates, none pertains to what are universally admitted to being the 
most archaic Yamnaya burials. At present, we can only surmise that there was a 
very rapid expansion of distinct ceramic types and burial ritual over a vast area. 
Nikolai Merpert has suggested that the almost instantaneous spread of 
elements so closely associated with expressing ethnic identity - ceramic style 
and burial ritual - may indicate the existence of substantial tribal unions 
engaged in intense contacts with one another. This would certainly be in 
accordance with our image of the Pontic-Caspian region as an enormous sphere 
of continual interaction and mutual influences, in which cultural traits and 
human groups traversed with great rapidity. If the inhabitants of the regions 
also shared a broadly similar language at this time, this would no doubt have 
assisted in the rapid diffusion of common cultural traits, and the creation of a 
broadly similar cultural horizon. 

Proto-Indo-European culture 

Any attempt to compare the linguistic evidence for Proto-Indo-European 
culture with the archaeological evidence of our supposed homeland region 
invites a certain dread among those with more than a passing familiarity with 
the literature on the subject. Precisely the same evidence employed to paint the 
Indo-Europeans into a Baltic landscape has been used to set them securely in 
the ancient Bactria of northern Afghanistan. There is indeed a curious tendency 
among some enthusiasts to assume that when the Proto-Indo-European 
vocabulary refers to clouds, rain and ice, these elements were somehow 
specifically abusing the inhabitants of their own particular homeland and 
nowhere else. Consequently, it must be admitted at the outset that the great 



majority of items concerning Proto-Indo-European environment or culture can 
hardly be ascribed exclusively to the Pontic-Caspian region rather than to 
elsewhere. On the other hand, if we do assume that the Pontic-Caspian region 
was inhabited by Proto-Indo-European speakers, then we can hardly avoid 
evaluating the extent to which the linguistic and archaeological evidence 

As expected, the Pontic-Caspian offers plentiful referents for the admittedly 
generic terms for the landscape which we can reconstruct linguistically. There 
are ample plains and rivers and other natural features. The problematic Proto- 
Indo-European *mori- be it an inland sea, salt lake or marsh, could still be 
accommodated within the Pontic-Caspian. Here there were several major 
bodies of water plus numerous deltas, marshes and now dried lakes that once 
served as primary settlement locations during the Neolithic-Eneolithic. 

The Proto-Indo-European vocabulary clearly attests words for mountains 
and high places. This fact has been pressed by the Soviet linguists Tomas 
Gamkrelidze and Vyachislav Ivanov to support their hypothesis that the 
homeland was in a mountainous region (Armenia). They dismiss as potential 
homelands the 'flatlands encompassing the whole of eastern Europe, including 
the Pontic'. As we have already observed, this surely raises the linguistic 
evidence to literally too great a height, since one hardly need live on a mountain 
to have a name for it. It is true that most of the mountain ranges such as the 
Carpathians, Crimean, Caucasus and Urals are largely peripheral to the Pontic- 
Caspian cultures, but we should never forget that geographical terms are 
merely perceptual categories. There is quite obvious relief within the Pontic- 
Caspian, and we should surely be astonished if its inhabitants actually lacked a 
word for mountain or hill. 

The climatic evidence is vague. The current conditions of the Pontic- 
Caspian vary in aridity as one moves eastwards and it is valid to characterize 
even the western part as possessing a moderate and somewhat arid climate with 
hot summers and short, mild winters. Currently the Pontic-Caspian sees about 
forty to eighty days of snow lying per year, and the conditions during the 
Eneolithic are assumed to have been drier but colder than at present. 

From the arboreal point of view, the Pontic-Caspian offers a number of 
diverse environments. The forest-steppe and the river valleys could provide all 
the necessary arboreal referents, while the steppe would have contributed little, 
and the semi-desert of the Caspian almost nothing. All the strongly attested tree 
names do find their counterparts in the Pontic-Caspian during the Neolithic- 
Eneolithic. It must be admitted that the beech, the most debatable arboreal 
item, did have a much more limited distribution. The common beech was 
barely found east of the Dnieper, while the eastern beech was confined to the 
Caucasus and Crimea. In such a situation one could hardly be surprised if the 
word is attested only in the European languages and does not find a reflex in the 
more easterly Indo-Iranian languages. 

The wild fauna reconstructed linguistically can be correlated with remains 
recovered from archaeological sites, although they are not unique to our area. 



Both the Indo-European vocabulary and the archaeological evidence empha- 
size a variety of riverine fauna such as beaver and otter, and one should not 
imagine the Pontic-Caspian purely as an arid steppe region. There are, of 
course, some animals known from the steppe sites, such as onager and saiga 
antelope, that are in no way reconstructible to Proto-Indo-European, though 
this is hardly surprising given their extremely limited distribution. 

The economy which we envisage from linguistic evidence for the Proto- 
Indo-Europeans is well-enough mirrored in the archaeological evidence. 
Agriculture is attested, but nowhere does it seem to have been the primary 
subsistence component of the economy. Indeed, in the Volga-Ural steppe, and 
possibly also in areas of the western steppe, there are reasons to exclude 
agriculture as a serious factor in the economy. Rather, the emphasis of the 
archaeological remains appears to be primarily on stockbreeding. Evidence has 
been found for all the basic livestock. There are no grounds to prefer sheep or 
cattle as the primary species since they vary with the particular ecological 
situation of each settlement. The domestic pig, a controversial animal among 
linguists, does appear to be of secondary importance throughout the Pontic- 
Caspian. It is attested minimally in the west and seems, so far, to have been 
absent in the east prior to the second millennium when it does appear on Bronze 
Age sites. This is coincidental with fairly intense relations running from the 
south to the north. As we have mentioned before, the Uralic languages acquired 
their name for the domestic pig from Indo-Iranians, and it is attractive to 
imagine this loan occurred here in the Volga-Ural region during the second 
millennium bc when Indo-Iranians may well have been in contact with their 
Uralic neighbours. 

The most significant linguistic-archaeological correlation among the domes- 
tic animals is the horse, which is known at least from the fourth millennium b c, 
or earlier from the Dnieper east to the southern Urals. It is not simply present, 
but all the evidence indicates intense exploitation and a role in ritual. All of this 
helps underwrite the apparently equo-centric practices that are shared by many 
Indo-European peoples. Where the horse occurs either to the south or 
immediately west of the Pontic-Caspian, this is generally explained by contacts 
with the natural home of the horse. This strengthens the Pontic-Caspian's 
claim to homeland status though it does not entirely secure it against other 
areas. Domestic horse remains are reported from both Central Europe in the 
Altheim culture and the North European TRB culture, but it should be 
emphasized that there is no evidence that these cultures exploited the horse in 
their economies or ritual on the scale we find in the Pontic-Caspian. 

The reconstructed vocabulary for settlement architecture is imprecise, and it 
is impossible to know whether the Proto-Indo-European *weik- y which spans 
such meanings as house, village or clan, might appropriately be assigned to a 
small settlement such as Dereivka. The postulated evidence for some form of 
fortified enclosure on the basis of Sanskrit, Greek, Armenian and Lithuanian 
can at least find a plausible referent in sites such as Mikhaylovka which are 
known from about 3000 bc. 



The basic assortment of Neolithic-Eneolithic technology evidenced from 
language is also mirrored on Pontic-Caspian sites. Obviously, little can be made 
of the fact that both the Proto-Indo-European lexicon and these sites exhibit, 
for example, pottery, grinding stones and arrowheads. The Proto-Indo- 
European thrusting (?) weapon, indicated by the Sanskrit-Latin cognates for 
'sword', may find its ultimate origin in either the long flint blades from the 
Sredny Stog and Novodanilovka cultures, the flint daggers known in the 
somewhat more recent Yamnaya burials, or in the earliest bronze daggers, 
w hich we will encounter in the next chapter. Copper is well attested and could 
offer the probable referent of the Indo-European basic metal term (*ayes-) and, 
as we will see below, silver is also known in the Pontic-Caspian from about the 
fourth millennium bc. If one accepts the linguistic existence of a Proto-Indo- 
European word (*(a)rgentom) for this metal, the Pontic-Caspian offers an 
especially attractive fit for a metal conspicuously absent from most of Europe at 
this time. 

The technology of the 'Secondary Products Revolution', which involves the 
plough, milking and other dairy products, wool, and wheeled vehicles, is all 
clearly indicated in the Proto-Indo-European vocabulary. The actual archaeo- 
logical evidence has seldom been assembled for discussion of this topic - faunal 
reports rarely explore the precise patterns of exploitation of livestock. The 
apparent increase in numbers of sheep, at least among communities exploiting 
the open steppe, may be associated with the words for wool and combing in the 
Indo-European languages. Indeed, Sandor Bokonyi has recently suggested that 
a new type of sheep, which was larger and offered more wool, had been 
introduced into the steppe region from Asia. Instruments putatively employed 
in assisting in milking cattle have been found on Yamnaya and later sites, and a 
primitive plough is recorded from Mikhaylovka. Only with wheeled vehicles 
and animal traction are we on really well-discussed ground, and there is 
abundant evidence for carts and wagons from at least 3000 bc onwards. This 
again is an attractive link with the linguistic evidence at a suitably early date 
although it is still not a unique correlation. Models of wagons are known from 
the Baden culture in Hungary at about this same time, and a four-wheeled 
wagon is depicted on a pot from a TRB site in Poland. 

Although there has been a vast amount of literature on the burial remains and 
ritual of the Pontic-Caspian Eneolithic cultures, we are still far from being able 
to pronounce authoritatively on the social structure of their communities, nor 
are we able to make apt comparisons with our linguistic evidence. Burials may 
range widely in wealth but it is difficult to postulate evidence for a Proto-Indo- 
European 'protector', 'man with charisma' or 'master of the clan' on this 
evidence. One of the problems here is the nature of our mortuary evidence. 
Kurgans frequently contain multiple burials, and we are seldom certain that 
our sample of burials derives from the same general period. There is no 
question that some graves are markedly wealthier than others, but if we cannot 
compare them confidently with contemporary burials, we have no solid basis 
for evaluating the structure of Pontic-Caspian society in detail. We can only 



note, as observed by Alexander Hausler and others, that very often the primary 
burial in a kurgan belongs to a male. Indeed, Maryana Khlobystina, in 
examining the distribution of burials by sex on the Volga, noted that the 
overwhelming majority belonged to males and children. Consequently, she 
suggested that females were buried elsewhere, probably in not-yet-discovered 
flat graves, and that the emphasis on male burials indicated the strong 
patriarchal nature of steppe society. 

Given the nature of our evidence, one can imagine how elusive Georges 
DumeziPs three classes might be in the archaeological record. Elena Kuzmina 
suggests that those burials accompanied by horse remains might claim to 
represent the nobility postulated by the Dumezilian school. It we look at an 
isolated case, such as Khutor Khryashchevskogo I, we certainly find an 
attractive candidate for a royal figure. In this late steppe burial, the deceased 
was accompanied by a horse skull, several pots, a metal spearhead, several 
arrows, a mace and a stone hammer. But this burial is hardly typical of all horse- 
associated burials, many of which are accompanied by only additional animal 
remains, or nothing else. 

Evidence for age-sets is slight. In his compendious survey of steppe burials, 
Alexander Hausler notes how some kurgans appear to be given over primarily 
to children's burials. In addition, he notes a certain tendency for older males to 
correlate with a southwesterly orientation. We should also emphasize that 
relatively wealthy burials of children are quite common, and there are obviously 
important issues concerning the ascription of wealth and status among the 
Pontic-Caspian populations; these warrant closer study. As we have already 
mentioned, the burial rites which follow the Dnieper-Donets culture do exhibit 
a shift from collective burials in large pits to smaller group pits, or individual 
burials arranged in small groups within a larger cemetery. Some Soviet 
archaeologists have interpreted this shift as a structural reordering of society 
that sees the emergence of family or clan burial plots. 

We may hope to find evidence of Indo-European ritual behaviour on the 
anthropomorphic stone stelae and in the remains of animal sacrifices. It is not 
difficult, for example, to read into those stelae which depict male figures 
bearing battle-axes the representation of a sky-god, thunder-god or, more 
abstractly, the warrior function. Naturally, given the ease with which we can 
squeeze three of our postulated Proto-Indo-European divinities out of the same 
block of stone, our actual ability to make a credible association between the 
iconographic evidence and Indo-European myth is clearly suspect. It is 
remarkable that, given the iconographic detail on some of the stelae, this source 
of information has been largely bypassed by those most involved in 
reconstructing Proto-Indo-European religion. We might emphasize that the 
stelae do not exclusively depict male figures wielding axes, but a number 
portray a female figure apparently wearing a necklace of beads. Other motifs 
include what appear to be two large footprints, and stylized trees. 

There is abundant evidence for the deposition of animal remains in graves, 
and also for special sacrificial areas. In a rough survey of animal remains in 



tombs across the Pontic-Caspian, we generally find sheep as the most frequent 
offering, followed by cattle, and then horse. We may wish to read these figures 
as inversely proportional to the status of the offering, keeping in mind the 
victims sacrificed to the Iranian goddess Aradvi (100 horses, 1,000 cattle, 
10,000 sheep). On the other hand, Nikolai Merpert suggests that the paucity of 
cattle and horse remains compared with those of sheep may have had more to do 
with economic factors, since the steppe tribes were perhaps less inclined to 
waste valuable traction animals on ritual occasions. 

Some form of horse sacrifice is, at least according to both the ritual and 
mythic evidence, one of the expected patterns of ritual behaviour which we 
might predict for the Proto-Indo-Europeans. We have already seen the 
sacrifice of the stallion and dogs at the Sredny Stog site of Dereivka, and 
archaeologists have been quick to link this ritual deposit with the type of horse 
sacrifices we later encounter in India. If we expand our attention to the ritual 
sacrifice of three different species, one of which is the horse, then there are some 
examples which would not be out of place in an Indo-European 'Afunctional' 
ritual. At Grushevka a 'ritual' pit was uncovered which contained the bones of 
horse, cattle and sheep, while the same three species were also found in a burial 
at Gerasimovka I. These are exceptions, however, rather than patterns, since 
normally the remains of horse - often the skull or forelegs - is either found alone 
or accompanied by only one other species such as sheep or cattle. 

We have already mentioned the problem of establishing some form of 
explanation for the divine twins so often encountered in Indo-European 
mythology. Mortuary evidence offers us the Yamnaya burial at Novoalekseevka 
6/14, where the skeleton of a one-year-old child was found lying between two 
horse skulls. This is an exceptional burial, and there really is very little ev idence 



to postulate the twinning or doubling of horses, in draft or ritual, among the 
Pontic-Caspian tribes of the Eneolithic. The only other explicitly ritual context 
is from the cemetery at Sezzhee. Here, above the group of richest burials, the 
excavators discovered a ritual area which included the skulls and legs of two 
horses, as well as two pots, a harpoon and shell beads, all sprinkled with red 
ochre. At Sezzhee we have also the broken remains of a figure purportedly 
depicting a two-headed horse although two-headed cattle figures are also 
known from the site. Two horses are depicted on the lower register of the 
Kernosovka stela. 

Finally, one further element of equine ritual may be seen in the horse remains 
from Dereivka. In her analysis of the faunal evidence from the site, Valentina 
Bibikova observed that of the eighteen whole metacarpals recovered, seventeen 
of them were from the left leg. From this we may postulate some form of ritual 
prohibition against breaking the left foreleg of the horse which may lead us back 
to the right-left dichotomy so often observed among the Indo-Europeans. 

Provided that our expectations do not demand precise detail, most of the 
archaeological evidence from the Pontic-Caspian does make a reasonably solid 
fit with our reconstruction of Proto-Indo-European culture. In some areas, 
such as those pertaining to the domestic horse and wheeled vehicles, the fit is 
particularly striking, and we cannot find such close correspondences among 
many other Eneolithic cultures of Europe. None of this, how ever, is sufficient 
to impell us to claim the Pontic-Caspian as the exclusive homeland of the Proto- 
Indo-Europeans. This can only be acknowledged if the proponents of a Pontic- 
Caspian homeland can demonstrate expansions from this region which account 
for the dispersion of the Indo-European languages. 



Indo-European Expansions 

The Aryans left their homes ... on the ist of 
March. This settles the question of the climate 
of their original home. Had their homes been 
situated in a moderate zone, the Aryans would 
never, of their own free will, have made their 
exodus so early; they would have delayed it, if 
not until May, at any rate until the middle of 

Rudolph von Ihering, 1897 

The preceding chapter has provisionally assigned the earliest Indo-European 
speakers to the Pontic-Caspian region. As this territory encompasses well over 
300,000 square kilometres, it indicates an area whose size is certainly 
comparable to the territory of a major linguistic entity. We are therefore 
reluctant to assign additional territory to this 'homeland' unless compelled to 
do so by other evidence. But we can only avoid enlarging our prospective 
homeland area if we can actually trace an expansion from this region, carrying 
us into all of those territories where we subsequently encounter the historical 
Indo-Europeans. If this can be accomplished we are afforded with a solution to 
the homeland problem expressed at the right order of magnitude and not so 
large as to be linguistically implausible or geographically meaningless. Many 
scholars, both archaeologists and linguists, do support the notion that the 
Pontic-Caspian region defines the Indo-European homeland. Now we must 
evaluate whether we really can limit the Proto-Indo-Europeans to such a 
confined homeland and trace their expansions from it. 

If we demand absolutely conclusive archaeological evidence for folk 
movements from the Pontic-Caspian, and the subsequent absorption of 
peripheral regions by intruders from the steppe, then our task is hopeless. It is 
too much to expect that we now have all of the archaeological evidence at hand, 
or indeed that archaeologists even agree on the nature of the evidence required 
to demonstrate population movements, much less the linguistic absorption of 
native populations by intruders. Rather, we can do little more than trace 
plausible trajectories of Pontic-Caspian peoples into a number of critical 
territories. We need to demonstrate that peoples from the steppe expanded into 
Southeastern Europe to effect the formation of the immediate ancestors of the 
Indo-European peoples of the Balkans and Greece, and of those languages of 
Anatolia whose origins have been directly linked to the Balkans, such as 
Phrygian, Armenian, and possibly the Anatolian languages themselves. We are 



also required to establish a genetic link that seriously involves the Pontic- 
Caspian in the formation of the Corded Ware horizon of Northern and Central 
Europe, since it provides the presumed staging area for the later emergence of 
the Celts, Germans, Baits, Slavs and possibly other Indo-European groups. 
And, finally, we must discover traces of Pontic-Caspian expansions into Asia to 
explain the creation of the Tocharians and Indo-Iranians. As we began our 
survey, in Chapter Two, with the Indo-Europeans of Asia, we will examine 
this region first. 

Expansion into Asia 

Although many details may be disputed, there is most certainly a serious case 
for an expansion of peoples from the Pontic-Caspian into the steppe and forest- 
steppe east of the Urals. These movements appear to have begun as early as the 
fourth millennium bc, and they may have continued over several millennia by 
which time the major patterns of migration in this region were dramatically 
reversed by Turkic-speaking peoples such as the Huns. 

The easternmost culture to claim a relationship with the Pontic-Caspian is 
the Afanasievo whose remains are primarily confined to the Minusinsk Basin 
and the Altai. Burials provide the most frequent evidence - over 400 graves 
from approximately fifty cemeteries - but about ten settlements are also known. 
The distribution of these sites indicates that the Afanasievo people exploited 
both riverine territories as well as the deep steppe. The settlements, at least 
those on the Yenisey, are insubstantial, and are commonly interpreted as the 
seasonal camps of pastoralists who raised cattle, sheep/goat and, significantly, 
domestic horse. Although wheeled vehicles are not directly attested, archaeolo- 
gists suspect that the Afanasievans possessed them. The cultural successor to 
the Afanasievo culture, the Okunevo culture, does exhibit evidence for wheeled 
vehicles, despite the fact that its economy appears originally to have been based 
on hunting and gathering. Consequently, we attribute the introduction of 
wheeled vehicles among the Okunevo people to their immediate and 
technologically more advanced predecessors. 

The Afanasievo dead were interred in rectangular pits which were covered 
by kurgans and surrounded by cromlechs and rectangular stone enclosures. 
The deceased were buried flexed on their side or, more notably, on their backs 
with their legs flexed, the traditional burial posture known from the Pontic- 
Caspian. Ochre was included in burials. The grave goods included pointed 
based pots, similar in general form to those of the Pontic-Caspian, and 
decorated with stamped impressions. Indeed, Maryana Khlobystina has 
synchronized the Afanasievo ceramics from the Altai with parallel develop- 
ments of Yamnaya ware on the Volga. Censers decorated with ochre were part 
of the Afanasievo ritual paraphernalia. Formally, they closely resemble similar 
cult artefacts known from the Pontic-Caspian. In addition to the predictable 
stone and bone tools found in the graves, there were also objects of copper, gold 
and silver. Faunal remains of sheep, and more rarely of cattle and horse, were 



i2j, 124 Afanasievo kurgan and burials, together with pottery and a censer (tallest vessel 
c. 25 cm). 



found in the fill of the burials and were interpreted as the residue of funeral 
feasts. Khlobystina notes that, like the earlier Yamnaya burials on the Volga, 
the earliest Afanasievo tombs in the Altai were confined exclusively to males 
and children. 

In contrast to the physical type of their neighbours, the Afanasievo people 
have repeatedly been classified as Europoids by Soviet anthropologists, who 
find their closest parallels with the Yamnaya populations of the Pontic- 
Caspian. Traditionally, Soviet archaeologists have dated the Afanasievo 
culture to the mid-third millennium bc; however, radiocarbon dates indicate 
that this culture probably began before 3000 bc and then continued through- 
out much of the third millennium. 

From the above, it is not difficult to understand why archaeologists have 
often associated the Afanasievo culture with the Pontic-Caspian. Similarities in 
burial ritual, material culture, economy and even in physical type all point 
towards the Pontic-Caspian. Moreover, if Elga Badetskaya is correct when she 
finds no solid evidence for stockbreeding on the Yenisey prior to the Afanasievo 
culture, then we have a hint of an explanation as to how this culture managed to 
spread so far to the east. Possessed with an already-developed mobile economy 
based on stockbreeding, populations from the Pontic-Caspian could swiftly 
have expanded to exploit the vast steppe and forest-steppe east of the Urals. 
Such an expansion would, no doubt, have brought the ancestors of the 
Afanasievo culture into contact with the various local sub-Neolithic communi- 
ties whom they eventually assimilated. This model seems plausible enough but 
it has always faced one major obstacle - the distance between the Ural, the 
traditional eastern limit of the Pontic-Caspian cultures, and the Yenisey, is in 
the order of 2,000 kilometres! 

The enormous gap between the Pontic-Caspian and the cultural territory of 
the Afanasievans is too obvious a difficulty to be ignored. This gap, however, 
may have more to do with the paucity of archaeological exploration in the 
intervening territory than to a genuine absence of intermediate sites. As our 
archaeological evidence to both the east and north of the Afanasievo culture 
reveals only hunting-gathering peoples, an Afanasievo origin somewhere to the 
west is all that we can seriously entertain. Recent discoveries now suggest the 
existence of intermediate sites that might help fill the gap between the Ural and 
the Yenisey. Tamilla Potemkhina, for example, has uncovered cemeteries at 

725 The region between the Yamnaya and Afanasievo cultures. 



Verkhnyaya Alabuga and Ubagan I in the Tobol drainage east of the Ural. The 
graves included burials in the supine position with legs flexed, the use of ochre, 
bone plates similar to those recovered from the Khvalynsk cemetery of the 
Volga, pointed-based pots similar to both Yamnaya and Afanasievo vessels, and 
copper artefacts provenanced by spectrum analysis to the Yamnaya region. 
These burials were in flat graves typologically similar to the very earliest 
Yamnaya burials of the Volga-Ural group. While Potemkhina observes that 
these burials indicate that the Yamnaya culture extended further east than had 
previously been suspected, Elga Badetskaya maintains that, had the ceramics 
from these sites been found on the Yenisey or the Altai, they would probably 
have been ascribed to the Afanasievo culture. Further east of the Tobol, on the 
Karaganda, a burial has been uncovered which was again in the classic 
Yamnaya posture and accompanied by a vessel similar to Afanasievo ware. In 
short, the evidence is slowly accumulating to support the existence of a vast 
extension of material culture, economy, ritual behaviour and physical type from 
the Pontic-Caspian eastwards to the Yenisey by about 3000 BC. 

To suggest an extensive band of communities historically related to the 
Pontic-Caspian across the West Asiatic steppe is of obvious linguistic interest. 
As we have seen above, some linguists would require for Tocharian origins a 
model which would place their ancestors on the 'archaic' periphery of the 
Proto-Indo-European dialects. This would explain why they did not partici- 
pate in the series of linguistic innovations experienced by their nearest 
historical neighbours, the Indo-Iranians. In purely temporal terms, an 
expansion of Indo-European speakers as far east as the Yenisey as early as the 
fourth or early third millennium BC might adequately account for the early 
separation of the ancestors of the Tocharians from their other Indo-European 
relations. It does not, of course, place them specifically in their later historical 
position in the Tarim Basin, which lies nearly 1 ,000 kilometres to the south. This 
distance, however, is not insurmountable when one considers that we have 
several thousand years to account for the movement of these apparently mobile 
Indo-Europeans either south of the Altai or through Mongolia into Chinese 
Turkestan. Indeed, long ago Sergei Kiselev suggested that the border of the 
Afansievo culture may well have extended further south than is presently 
known. Kiselev referred to the discovery of sherds of pottery, with decoration 
similar to Afanasievo ware, recovered by Sir Aurel Stein in the western part of 
Chinese Turkestan, the area immediately adjacent to the earliest-attested 
Tocharian-speaking people. 

Recent excavations have also suggested that the trajectory of Pontic-Caspian 
movements was not only to the east but also towards the south where steppe 
elements have been found in the fourth- or third-millennium BC cemetery of 
Tumek-Kichidzhik. The site is located south of the Aral Sea and is assigned to 
the Kelteminar culture, the major Neolithic (but with a hunting-gathering- 
fishing economy) culture of the east and southeast Caspian. As well as clearly 
local traits such as lithics and ceramics, the grave goods included ornaments of 
bone and boar tusk which were nearly identical to goods found in the Samara 



and later cultures on the middle Volga. In addition, the burial rite - extended 
supine position, northern orientation, ochre - and the proto-Europoid physical 
type also relate to the steppe. Alexandr Vinogradov links these steppe 
influences with the southern movement of sheep-rearing pastoralists who 
expanded to take advantage of the desert and semi-desert conditions of this 
region. This is a pattern that was maintained through the early Bronze Age 
where we find traces of steppe burials on the northern borders of the proto- 
urban settlements of Central Asia. 

The later Steppe Bronze Age of western Siberia provides us w ith our second 
major watershed that may be associated with the expansion of Indo-Europeans 
in Asia. Traditionally, the second millennium bc of the west Siberian steppe is 
primarily represented by the Andronovo culture. This, in fact, is a cultural label 
that embraces a series of local cultural groups which occupied the forest-steppe 
and steppe from the Urals to the Yenisey and from the northern border of the 
forest-steppe south to the Pamirs of Tadzhikistan. Admitting the existence of 
predictable regional variation, this enormous region does find similar ceramics 
and metal types, a predominantly stockbreeding economy, and a range of 
broadly similar mortuary practices and ritual traditions. Today the concept of a 
unified Andronovo culture has been seriously challenged by a number of Soviet 
archaeologists who prefer to regard the regional variants as independent 
cultures existing within the context of a vast sphere of interactions that 
generated the similarities mentioned above. Radiocarbon evidence suggests 
that the Andronovo culture may have begun to emerge in the early second 
millennium bc. 

Few, if any, archaeologists would deny a general Indo-Iranian identity for 
most of the bearers of the Andronovo culture, nor would they deny its 
fundamental genetic association with its western neighbours in the Pontic- 
Caspian. These connections are not only of the more general kind - primarily 

126 Generalized distribution 
of the Andronovo and 
Srubnuyu cultures. 



stockbreeding economy, domestic horse, wheeled vehicles, kurgan burials - but 
they also involve the specifics of ceramic form and decoration, as well as the 
types of metal implements and ornaments. Indeed, the Pontic-Caspian 
successor to the Yamnaya culture, the Bronze Age Srubnaya (Timber-grave) 
culture, appears to have penetrated east of the Volga, and Srubnaya burials are 
known from the east Caspian. 

The Indo-Iranian identity of the Andronovans is founded on both cultural 
and geographical evidence. The cultural remains from the Steppe Bronze Age 
sites coincide with our expectations for the immediate ancestors of those Indo- 
Iranians whom we later encounter in the historical record. The terrain of the 
region was one normally selected for a mobile economy founded on 
stockbreeding, and with few exceptions this is what we encounter. The herd 
consisted of cattle, sheep/goat and horse, and there appears little or no place for 
the domestic pig, a situation predicted by the poor lexical preservation of the 
inherited Indo-European word in the Indo-Iranian languages east of the Urals. 
The horse was clearly domesticated (it accounts for from 1 2 to 27 per cent of the 
domestic fauna), and was employed both in traction as well as being ridden: 
horse psalia, designed to hold the bit in the horse's mouth, are found on 
Andronovo sites, and paired horse burials suggest its use in drawing vehicles. A 
clay model of a horse with a perforation through the mane, putatively for 
attachment to a vehicle, is known from one of the southernmost Andronovo 
variants. More spectacular evidence for the chariot-borne aristocracy encoun- 
tered in the later historical record is the major cemetery at Sintashta in the south 
Ural steppe. Here the remains of chariots and sacrificed horses clearly presage 
the later royal burials of the steppe tribes such as the Scythians, or their more 
easterly Iranian cousins, who buried their dead in the spectacular chambers 
recovered from the Pazyryk cemetery of the Altai. Vladimir Gening, the 
excavator of Sintashta, has utilized Indie and Iranian religious literature as a 
key to interpreting the rituals evident in the cemetery. He calls attention to the 
ritual horse sacrifice (with as many as seven horses in a single burial chamber), 
dog sacrifice and the slaughter of domestic animals. Here also we might have a 
hint of the type of trifunctional animal sacrifices known from early Indo- 
European literature. In pit 1 was discovered a sacrificial area apparently 
unassociated with burials. Within it were the remains (especially skulls) of 
horses, bulls and sheep. 

Unlike earlier essentially pastoralist peoples, the Andronovans have left us 
with considerable remains of their settlements which were generally situated 
along small rivers and ranged from two to twenty houses. These were 
constructed of wood - pine, cedar and especially birch which, as we noted 
earlier, is one of the very few Proto-Indo-European arboreal names well- 
preserved in Indo-Iranian. Elena Kuzmina has suggested that settlement 
evidence may be used to identify the Andronovans with the Indo-Iranians. She 
emphasizes the differences between the dwellings of both the Indus civilization 
and the proto-urban communities of Central Asia, and those of the steppe 
pastoralists. The former involved brick and stone surface structures with small 



rectangular rooms in the order of 9 to 12 square metres. These, she argues, 
correlate with small family units and are in striking contrast to the semi- 
subterranean timber-built houses of the Andronovans which occupied areas 
from 50 to 300 square metres, dimensions suggestive of large extended families. 
The evidence from both the Vedas and the Avesta indicate that the Indo- 
Iranian house can only be paralleled by those of the steppe and not by the 
indigenous (proto-)urban settlements. 

Further support for identifying the Andronovans with the Indo-Iranians is 
that all our evidence for the trajectory of the earliest Indo-Iranians is from the 
north to the south, and consequently the steppe is the most logical staging area 
for their later migrations. Finally, the territory of the Andronovo culture 
coincides with the later Iron Age territories of the historically attested Eastern 
Iranians - the Saka, Massagetae, Sarmatians, and Alans. 

Although the Steppe Bronze Age peoples make an admirable fit with the 
territory and expected culture of the steppe Iranians, it is not easy to make a 
simple appeal to the Andronovo culture to resolve all the issues of Indo-Iranian 
origins. Neither Western Iranians putatively associated with grey wares, nor 
the Mitanni nor Indo-Aryans, and the variety of their possible congeners in the 
Swat Valley or Indus region, are closely linked to the Andronovo culture of the 

When the archaeological evidence becomes so opaque then our only refuge, if 
we choose to pursue this problem further, is probability, and not a little 
intuition. We can, I think, accept as one anchor the assumption that the Late 
Bronze Age cultures of the steppe were, for the most part, the linear ancestors of 
at least the Eastern Iranians of the historical period. Once this is admitted, then 
where is the most likely location for the Western Iranians and the Indo-Aryans 
during the Bronze Age? It is reasonably safe to exclude them from land east of 
the Yenisey or north of the forest-steppe. It is always possible that some 
Iranians were situated to the west of the Andronovans in the Srubnaya culture 
of the Pontic-Caspian, and we have already reviewed the suggestion that these 
may have passed southwards through the Caucasus to form the Western 
Iranians. But as for the ancestors of the earliest Indo-Aryans (and, I suspect, 
Western Iranians since the Caucasian theory leaves much to be desired), it is 
most probable to seek them along the southern border of the Asiatic steppe 
during the second millennium b c, if not earlier. If they are not clearly evident in 
the archaeological record this is probably due to the nature of processes 
involved in any movement south from the steppe. In general, archaeologists 
concerned with the problem of Indo-Aryan or Western Iranian origins believe 
that it entailed gradual and protracted movements from the steppe through 
Central Asia. Such a movement would require the mobile populations from the 
Eurasian steppe to pass through a perimeter of the more advanced agricultural 
societies of Soviet Central Asia and northern Afghanistan. This would not only 
include intimate contacts with proto-urban towns, but even with outposts of 
the Harappan culture, such as Shortugai. Such contacts would then transform 
the material culture of the steppe pastoralists, especially in the spheres of 



i2j Many associate the Audrtnove 
culture with the early Indo-Iranians. 
The southernmost Andronovo tnhes 
would have had to pass through the 
cultural area of Central Asian 
urban ism and outposts of the Indus 
I alley culture (Shortugai) before 
arriving in the area of the Gandhara 
Grave culture (Swat Valley) or 
northwest Indus. 

ceramics, architecture and metallurgy. For this reason, we should hardly expect 
to find our archetypal Andronovo pastoralists south of the steppe itself. 

The process of cultural transformation outlined above is by no means 
hypothetical. South of the Aral Sea, and along the Amu Darya, there existed the 
Tazabagyab culture, frequently regarded as a local variant of the Andronovo 
culture. In addition to its livestock, including the horse, we also encounter 
irrigation systems along which small Tazabagyab villages, consisting of perhaps 
ten families, began to appear. We should also note that Tazabagyab burials 
signalled the sex of the deceased by burying males on their right side and 
females on their left, a trait that is hardly unexpected among the Indo- 
Europeans and which is encountered elsewhere in the Andronovo cultural 
region, such as at Tulkhar. Further east along the Amu Darya we encounter 
other Andronovo variants in Uzbekistan and Tadzhikistan which were either in 
direct contact or partly related to the more 'advanced' agricultural populations 
immediately to their south, including northern Afghanistan. Nataliya Chlenova 
has recently emphasized the importance of the major waterways of the area as 
historical routes which carried people between Central Asia and the Indus 
Valley. She notes that the southernmost Andronovo burials occur alon^ the 
river Vaksh, w hich is less than 500 kilometres from the region of the putatively 
Indo-Aryan Gandhara Grave culture and may be reached by the river Kabul. 
Moreover, Elena Kuzmina perceives parallels between both the domestic 
architecture and burial rites of the Gandhara Grave culture, and the regional 
variants of the Andronovans to their north. 



It is perhaps here that probability must yield to intuition, at least until the 
volume of excavated material from the region between the Indus and the steppe 
has been considerably augmented. The existence of Indo-Aryans and Iranians 
south of the steppe directs our attention to the complicated transition zone 
between the steppe and the band of more settled agriculturalists which 
extended from the Caspian east to the Amu Darya Basin, and beyond. Already 
by the fourth or early third millennium, steppe pastoralists had come into 
contact with Central Asian agriculturalists. Later we find the Dashli and Sapalli 
Tepe cultures emerging in the second millennium bc. These show a 
combination of factors - the indigenous traditions of the Central Asian proto- 
urban settlements such as Namazga and Altin Depe, and influences from the 
Andronovans. Their ceramics provide some of the better analogies for the type 
of pottery we find in the Gandhara Grave culture. Elsewhere in Central Asia 
Andronovo pottery is found on proto-urban sites, and Elena Kuzmina and 
others suggest that the steppe pastoralists may have been a major factor in the 
collapse of these sites in the late second millennium. With the steppe already 
occupied by Indo-Europeans from Afanasievo times, we may speculate that 
subsequent Indo-European populations evolved into the Indo-Iranians. We 
find some evidence on the major approaches into both India and Iran that 
indicates the movements of mobile pastoralists, possessed of the basic 
characteristics we might expect from the earliest Indo-Iranians. The details of 
how they ultimately arrived in their historic homes are still unclear, but the 
connection between the Pontic-Caspian and their immediate staging area in the 
steppe seems reasonable enough. 

Expansion into the Caucasus 

In our survey of Anatolian origins, we have seen that there are two schools of 
thought on the earliest appearance of the Anatolian languages. Some prefer a 
western entry via the Balkans while others have opted for an intrusion that 
carried Indo-European speakers from the Pontic south through the Caucasus 
and into eastern Anatolia. Although the eastern entry seems to involve some 
critical problems, not the least being its inability to trace a migration as far as the 
historical seats of the Indo-European-speaking Luwians, it has had an 
attraction for some linguists. These have emphasized that the Anatolian 
languages share certain features with the languages of the Caucasus (or, like the 
Caucasian languages, lack certain features found in all other Indo-European 
languages). Moreover, a number of linguists have claimed that the Caucasian 
languages, such as Kartvelian, must have been in close contact with an Indo- 
European language, presumably an archaic one such as we might expect from 
the ancestors of Hittite and Luwian. 

An intrusion from the Pontic steppe into the Caucasus has been argued 
primarily by Marija Gimbutas. Her evidence is essentially founded on the 
appearance of kurgan burials in the steppe region of the confluence of the Kura 
and Araxes rivers. Among these kurgan burials are the three great mounds at 



Uch-Tepe, the third of which has been excavated. This enormous mound 
measured 1 7 metres in height and 1 30 metres in diameter, and enclosed a stone- 
built chamber roofed with timber beams which have been dated to about 3500 
bc. This, and other kurgans such as the one covering a wagon burial at Bedeni, 
have been interpreted as the tombs of the ruling elite who penetrated the local 
Kuro-Araxes culture from the Pontic-Caspian steppe. Other kurgans are 
known in western Georgia. In addition to what she sees as a kurgan burial rite, 
Marija Gimbutas also cites the settlement at Mingachaur as evidence of 
northern invaders. Here the rectangular semi-subterranean timber-built houses 
contrast with the typical Kuro-Araxes stone and clay-lined round houses. 
Finally, the appearance of Caucasian metallurgical types over the Pontic- 
Caspian is credited to intrusions by the Kurgan people into the Caucasus, who 
then transmitted the new techniques back to their kinsmen on the steppe. To 
this we might add the remains of horses, putatively domestic, from the site of 
Alikemektepesi in the steppe to the south of the Kura and Araxes. Soviet 
archaeologists cite this as the earliest evidence for domestic horse in the 
Caucasus, dated to the fourth millennium bc, if not earlier. Horse remains are 
regularly encountered on Kuro-Araxes sites. 

If we compare this evidence with that of the Pontic-Caspian, then the 
similarities involve the use of the tumulus, burial chamber, and in some cases 
such as Uch-Tepe, the use of ochre. In several more detailed considerations of 
this material, Shan Winn relates it to the later royal burials at Alaca Hiiyuk, 
Trialeti and Maykop, which brings the steppe, the Caucasus, and northern 
Anatolia into some form of contact relationship. Winn emphasizes that the heirs 
to the Kurgan intrusions of the Kura-Araxes region are unlikely to have been an 
historical Indo-European group but, more plausibly, Hurrians. Nevertheless, 
he does entertain the possibility that Indo-Europeans passed through the 
Caucasus to occupy eastern Anatolia whence they progressively assimilated the 
Hatti of central Anatolia to form the later Hittites. 

128 Kurgans in Transcaucasia. 



The evidence for a Pontic-Caspian intrusion into the Caucasus is based 
largely on the work of Soviet archaeologists such as Karine Kushnareva, Tariel 
Chubinishvili and Rauf Munchaev who do not themselves explain these kur- 
gan burials as the product of northern steppe intruders. Rather, they treat the 
barrows and the deposition of precious metals and wagons as a natural part of 
the process of social stratification witnessed throughout the Caucasus (the 
ceramics in the barrow graves are the same found in other forms of Kuro- 
Araxes burials). The Kuro-Araxes culture, with its relatively advanced 
metallurgy, distant exchange relations, and mixed farming subsistence basis 
could be expected to have evolved into a more ranked society which would have 
promoted the symbolic expressions of power and wealth. The social changes, 
therefore, can easily be dismissed as locally inspired, rather than intrusive, and 
there is no reason to deny to the ancestors of the Hurrians the same forms of 
display and social stratification which we might attribute to the Indo- 
Europeans. Similarly, exploitation of the steppe, and the creation of wealth- 
consuming pastoral societies in the Kuro-Araxes region, need not be specifically 
Indo-European, as numerous non-Indo-European populations throughout the 
Near East also evolved pastoral economies. We might also wonder about the 
antiquity of the Caucasian kurgans since the radiocarbon dates for Uch-Tepe 
are roughly contemporary with the earliest-dated kurgans on the Pontic- 
Caspian which are minute in comparison to this Caucasian tomb. 

The evidence for Pontic intrusions through the Caucasus is neither abundant 
nor easy to evaluate. We may emphasize, however, that where the evidence for 
barrows is found, it is precisely in regions which later demonstrate the presence 
of non-Indo-European populations; hence the model of Pontic intrusions 
through the Caucasus at this time may be largely irrelevant to the question of 
Indo-European expansions. 

Southeastern Europe and western Anatolia 

By 6000 bc, farming communities were progressively expanding through 
northern Greece into the Balkans. From their first appearance, until about 4000 
bc, they appeared to have enjoyed two millennia of uninterrupted develop- 
ment. This is seen in the expansion of population; the increasing density and 
size of settlements; the dev elopment of stone and early copper metallurgy; the 
refinement of subsistence strategies based on mixed farming; incipient craft 
specialization; elaborate ritual paraphernalia; and, probably, the evolution 
towards more complex social organization. Archaeologists witness this 
progressive development both in the correct ordering of the thousands of 
individual archaeological sites or, more conveniently, in a number of tells where 
the settlement debris of 2,000 years of continuous settlement has accumulated 
into massive mounds. The final phase of this development may be seen in the 
tells of south Bulgaria of the Karanovo VI period, the Gumelnitsa culture to its 
north, the late Vinca and contemporary cultures of the central and west Balkans 
and the Cucuteni-Tripolye culture on the western fringe of the Pontic region. It 



is during this final period of indigenous development, 4500-4000 bc, that 
archaeologists observe the earliest traces of direct contacts with the peoples of 
the Pontic-Caspian. In the subsequent millennia the evidence for contacts 
swells dramatically enough to convince archaeologists such as Vladimir 
Dumitrescu, Milutin Garasanin, Frano Prendi, Henrietta Todorova, and many 
others, that there was a sizeable influx of people from the Pontic region who laid 
the foundations for the Indo-European languages throughout Southeastern 
Europe. Some, such as Marija Gimbutas, would even credit them with the 
apparent collapse of two millennia of local development, and the total 
reorganization of the region under the aegis of Indo-European intruders. The 
evidence for this expansion of Pontic peoples is variable in quantity, complexity 
and credibility. 

The earliest ev idence for steppe peoples in Southeastern Europe is seen in 
the appearance of a series of interrelated objects and ritual behaviour that are 
without precedent in this area but which find close analogues in the Dnieper 
region during the Sredny Stog and Novodanilovka periods. The material 
evidence includes stone 'sceptres' fashioned in the shape of horse heads, 
typically encountered in the Pontic-Caspian but which are now also found in 
Southeastern Europe. Their total number exceeds thirty, and they may be 
divided into three types, largely on the degree to which they realize the actual 
features of the horse. The most realistic are mainly confined to the west of the 
Dnieper, with the exception of a sceptre from Terekli Mekteb in Dagestan. 
They are also found on Tripolye sites such as Fedele§eni in Romania, and east 
Balkan sites such as Salcuta and Casimcea in Romania, Rzhevo and Suvorovo in 
Bulgaria and Suvodol in southern Yugoslavia. It is widely accepted that these 
are objects ultimately of steppe inspiration, where the horse played an integral 
role in both economy and ritual and w here a tradition of manufacturing stone- 
carved animal figures extends back well into the Neolithic. The context of the 
Suvorovo find is especially interesting since it accompanied a male burial in a 

i2g Horse-head sceptres of the Casimcea-Suvorovo type: 1 Casimcea, 2 Terekli-Mekteb, 
J Salcuta, 4 Suvorova, 5 Fedelesetti, 6 Rzhevo. (The Suvorovo sceptre measures c. 17 cm 



i JO Some major Balkan sites 
showing early evidence for 
intrusions from the steppe. 
Sites with horse-head sceptres 
and burials with steppe 
features are indicated by dots; 
major settlements arc indicated 
with triangles. 

double grave. The male was provided with a sceptre, a pestle for grinding ochre, 
and long flint knives generically similar to those encountered in Novodanilovka 
burials. The accompanying burial, presumed to be female, wore a necklace of 
shell beads, again similar to those found in steppe burials and contrasting with 
the typical limestone beads of the native Eneolithic culture. 

Even more than burial goods, the appearance of a specifically steppe burial 
ritual in Southeastern Europe is generally regarded as the most solid evidence 
for the presence of steppe intruders. The cemetery at Decea Muresului in 
central Romania offers nearly twenty graves displaying links with the steppe. 
These included long flint knives, Unio shell beads and stone maceheads similar 
to those recovered in the Novodanilovka group. The burial rite included the use 
of ochre and, more importantly, the positioning of the dead in the supine 
position with legs flexed, the classic Pontic-Caspian ritual which is unparalleled 
in the local cultures of Southeast Europe. The westernmost example of this 
particular form of burial was found in a single grave at Csongrad-Kettoshalom 
in eastern Hungary where a male was buried in the 'steppe position' and 
accompanied by a long obsidian blade, small copper beads together with 
limestone and shell beads, and a fragment of ochre. This complex of burials and 
stone sceptres is generally synchronized with steppe cultures dating to the 
centuries around 4000 BC. Even before this period, there is some ceramic 
evidence for contacts between the steppe tribes and the more settled 
agriculturalists immediately to their west. 

Before 4000 bc, Tripolve sites begin to exhibit a ceramic that is widely 
known in archaeological literature as Cucuteni C ware. It is distinguished from 
the classic wares of the Balkans and northwest Pontic by the presence of crushed 
shell in the temper and by cord decoration, two features which typify the 
ceramics of the Pontic-Caspian steppe tribes. Such ware varies in quantity on 



Cucuteni-Tripolye settlements but may easily account for 10 per cent of all 
ceramic remains on some sites. The presence of either steppe pottery or, more 
usually, pottery fashioned according to techniques regularly employed on the 
steppe, is generally interpreted as evidence for direct contacts between steppe 
pastoralists and their more settled Cucuteni-Tripolye neighbours. Such 
ceramic contacts are hardly unexpected, as we have already seen how Pontic- 
Caspian sites have been dated by the occasional presence of Tripolye wares in 
steppe burials. 

Interpretations of this initial phase of Balkan-Pontic contacts vary, but even 
the minimal view admits that there is something more than distant trading 
relations involved. While some may be tempted to dismiss the evidence of stone 
sceptres as a widely distributed status item circulating among various local 
elites, an interpretation similar to that employed by some for the Bell Beakers 
the sudden appearance of an alien burial rite in Southeast Europe which is also 
markedly similar to that of the steppe - does suggest an intrusive population. 
But, though intrusive, some East European archaeologists have not regarded it 
as evidence for a substantial invasion. Instead, Istvan Ecsedy and others argue 
that what we are witnessing is a limited Pontic presence in Southeastern Europe 
which was conservative enough to retain certain behavioural features carried 
from the steppe but which was also active enough to acquire luxury goods such 
as copper and obsidian in its new home. 

The interrelationship between the settled agriculturalists of the northwest 
Pontic and the more mobile pastoralists of the steppe was not entirely 
symmetrical. Tripolye imports on steppe sites tend to be represented by well- 
made pottery often found within burials, clay figurines (again in burials) and 
copper objects. In a recent survey of such finds, Vladimir Zbenovich cites 
thirteen examples of Tripolye imports on steppe sites, frequently burials, 
between the Bug and Dnieper rivers. These imports, both by their nature and 
context, suggest prestige items acquired by the steppe pastoralists from their 
technologically more advanced neighbours. On the other hand, the various 
quantities of shell-tempered and cord-decorated wares that are found in 
increasing amounts on later Tripolye sites can hardly be interpreted as the 
result of an exchange between community leaders. Rather, like the occasional 
presence of antler cheekpieces on Tripolye sites, the steppe influences on their 
agricultural neighbours may be the result of direct contacts, possibly involving 
seasonal visits by Pontic-Caspian pastoralists to Tripolye settlements. This is 
also the pattern of behaviour attributed to Yamnaya herdsmen who may have 
gathered seasonally at major settlements such as Mikhaylovka. The relation- 
ship between the pastoralists and the settled farming communities may not 
have been purely symbiotic, since we find an increasing tendency towards 
larger, fortified settlements in the later Tripolye period. Maidanetskoe, for 
example, is an enormous Tripolye settlement with over 1,000 structures 
surrounded by several ditches. In human terms, then, the inhabitants of the 
large and usually fortified sites of the later Tripolye culture may have been 
forced to accustom themselves to the periodic visits, friendly or otherwise, of 



ijj Maidanetshoe is one of the largest late 132-134 Objects from the Usatevo culture. 
Tripolye villages. The surrounding (Top) painted ware of the Tnpolye 

fortifications are sometimes attributed to tradition; fright,/ dagger of arsenical 

defensive measures against steppe bronze; f above) cord-decorated coarse ware, 


their steppe neighbours. Ethnographic evidence suggests a very fluid boundary 
between mobile and settled communities, and it is entirely probable that some 
pastoralists may have settled permanently whilst Tripolyeans may have become 
integrated into the more mobile steppe communities. The resultant archaeo- 
logical evidence certainly suggests the creation of hybrid communities. 

By the middle of the fourth millennium bc we witness the transformation of 
Late Tripolye groups into new cultural entities. Probably the most noted is the 
Usatovo culture which occupied the territory from the lower Dniester to the 
mouth of the Danube. The available radiocarbon dates suggest that it 
flourished about 3500 to 3000 bc. In some aspects the culture retains traditional 
Tripolyean styles of painted wares arid figurines and, at least for the Late 
Tripolye period, flat cemeteries. But, in addition, there also appears a ceramic 
technology that is fundamentally characterized by shell-tempered wares and 
cord ornament, a lithic industry shaped by steppe influences, and the use of 
large kurgans for burials. Metal objects are relatively abundant and include 
arsenical bronzes. Among objects found were a considerable series of daggers, 
along with axes, awls and rings, including rings made from silver which is a 
metal we would attribute to the Proto-Indo-Europeans. Economically, the 
subsistence basis differs from the typical Tripolyean cattle- and pig-based 
economy to one where ovicaprids are dominant and horse varies to between 10 
and 1 5 per cent of the faunal remains. 



The creation of such a hybrid culture is not easy to explain, and Soviet 
archaeologists themselves maintain no pretence of consensus. Some would see 
the presence of kurgan burials, which were sometimes more richly furnished 
than the more traditional flat cemeteries, as evidence that steppe overlords had 
superimposed themselves on the local Tripolye population. This is probably far 
too simplistic an explanation, especially since Tripolye communities do not 
even appear to have occupied this territory prior to the emergence of the 
Usatovo culture. Instead, we must also take into account 1,000 years of 
population growth to the north of the Usatovo culture on the middle Dniester 
and southern Bug, and subsequent expansion southwards where the traditional 
Tripolye economy was far less suited. Movement into this area, some argue, 
brought the Tripolyean element into contact with steppe pastoralists and 
possibly with communities from the Caucasus, all of whom participated in the 
creation of the Usatovo culture. 

The apparent hybridization between steppe elements and formerly settled 
agricultural communities of Balkan origin is seen further to the west in the 
Cernavoda I culture of eastern Romania. Here, also, a coarse ware - similar to 
the Cucuteni C ware of Late Tripolye - appears, with its shell temper and cord 
decoration, together with the sporadic appearance of horse. This region, 
immediately south of the Cucuteni-Tripolye territory, was previously occupied 
by the Gumelnitsa (Karanovo VI) culture. During the fourth millennium, 
archaeologists perceive a structural reorganization of society across much of 
Southeastern Europe. Evidence for this comes from the abandonment of the 
tell sites which had flourished for several millennia; the displacement of 
previous cultures in almost every direction except eastwards; movement to 
marginal locations, such as islands and caves, or easily fortified hilltop sites such 
as Cernavoda I; and a general reduction in the major Eneolithic technologies of 
both fine ceramic manufacture and copper metallurgy. This abandonment and 
movement, often propelling neighbouring cultures into one another, operated 
against a background not only of somewhat elusive traces of hybridization with 
the steppe cultures such as the Usatovo and Cernavoda I, but also with 
continuous incursions of mobile pastoralists. 

The cultural chaos of this period has produced something of a Balkan 'dark 
age\ The tell sites, for example, number in the order of 600 to 700. Their 
continuous development comes to a halt when they are abandoned about 4200 
bc, and so far we only have evidence of several dozen being reoccupied during 
the subsequent Early Bronze Age, somewhere between 500 and 1,000 years 
later. Out of this period there later emerges a new cultural horizon that 
integrates cultures across Eastern Europe, including the northwest Pontic, and 
western Anatolia. Perhaps the most dramatic example of this new horizon is 
the settlement of Ezero in south Bulgaria. Here the accumulation of Neo- 
lithic and Eneolithic debris halts during the Karanovo VI period and then, 
after abandonment, we find the Early Bronze Age Ezero culture (Karanovo 
VII) beginning in the centuries around 3200 bc. On the ruins of the previous 
Eneolithic village was erected a settlement about 60 to 80 metres in diameter, 



*35 The circum-Pontic 
interaction sphere. 

surrounded by stone walls. In subsequent periods the Ezero walls were 
extended, and throughout the Early Bronze Age Ezero took on the appearance 
of a citadel, prov iding a defensive focus to small unfortified settlements in the 
vicinity. The Ezero culture shares numerous similarities in fine ceramics, 
plastic art and metallurgy with other emergent Early Bronze Age cultures 
throughout Southeast Europe, particularly with the Baden culture of the 
Danubian region and the Cofofeni culture of Romania. This new Balkan- 
Danubian complex of related cultures was not confined to Southeast Europe 
but now extended across into northwest Anatolia and the Troy culture. It is 
important to emphasize that at Ezero, the excavators Nikolai Merpert and 
Georgy Georgiev discovered a sequence of ceramics clearly ancestral to, and 
then contemporary with, those found in the Troy culture. The evidence may 
still be regarded as tenuous, but a growing number of archaeologists now admit 
the temporal and cultural priority of Southeast Europe to the Early Bronze Age 
of northwest Anatolia. Once this is accepted, there is no great step to 
entertaining seriously the proposition that an influx of people from the Balkans 
carried both the material culture and possibly Indo-European languages into 
northwest Anatolia in the centuries around 3000 bc. 

This consolidation in Southeast Europe was played out against a continuous 
background of further intrusions from the steppe. What was sporadically 
attested prior to 3000 bc swelled during the third millennium to provide 
unequivocal evidence for a movement of populations from the Pontic-Caspian 
steppe into the Balkans. If we take the lower southwest corner of the Soviet 
Union, the area between the Danube and the Dniester rivers, as our trip line, we 
find abundant evidence for the staging area of these highly mobile populations. 
In a recent study of the Eneolithic and Early Bronze Age kurgans of the region, 
Yevgeny Yarovoy lists over 1,000 excavated burials from some 118 sites in this 
region alone. Ukrainian archaeologists such as Ivan Chernyakov, Nikolai 
Shmagliy and Gennady Toshchev have examined the material classifying it 
into four periods: the pre-Usatovo burials w hich we have already referred to as 
evidence for the initial Pontic-Caspian contacts with Southeast Europe; the 
Usatovo kurgans; and the kurgan burials dating to the early and later Yamnaya 



rj6 The distribution of 
excavated king an* (dots) 
in the lower Danube. 

periods. The bulk of the remains date to this later period which takes us though 
the third millennium bc. 

The evidence for a westward movement of Pontic-Caspian peoples is not 
limited to the Danube; kurgan burials now appear in Romania, Bulgaria, 
Yugoslavia and as far west as the Tisza river in Hungary. These burials mirror 
the mortuary rite of the Pontic-Caspian with the supine flexed position, use of 
ochre, employment of w ooden mortuary structures or mats of reed or grass, and 
even the physical anthropology of the deceased speaks for an intrusive 
population. Studies on kurgan burials in Romania, for example, have revealed 
that the more robust-appearing kurgan males averaged up to 10 centimetres 
taller than the native Eneolithic population. Grave goods such as silver 
earrings, copper beads, and the knucklebones of sheep (for children), all similar 
to items found in steppe burials, are known from these Balkan and Hungarian 
tombs. We can gain a hint of the magnitude of these migrations when we note 



that Istvan Ecsedy reports over 3,000 kurgans in Hungary alone, although only 
about forty-five have seen excavation. 

A new cultural framework for Southeast Europe, erected against a backdrop 
of continuous population movements from the steppe, underwrites the 
historical connection between the Pontic-Caspian and Southeast Europe and 
Anatolia. It should be emphasized that, in many of the later Yamnaya burials, 
the ceramics recovered owe more to the cultures of the new configuration than 
to the steppe. These, and other features such as the stone-constructed citadels 
of Mikhaylovka, Ezero and Troy, have directed Nikolai Merpert's attention to a 
circum Pontic sphere of mutual interrelationships in architecture, ceramics and 
metallurgy in the Early Bronze Age. 

The model for the prehistory of the fourth and third millennia derived from 
archaeology can accommodate the expectations of the historical linguists. One 
might imagine, for example, that the Anatolian languages were introduced into 
northwest Anatolia by the end of the fourth millennium bc, in association with 
the movement ascribed to the Balkan-Danubian complex. This would separate 
them earlier from their other Indo-European cousins. The innovating core 
which links Greeks, Armenians, Iranians and Indo-Aryans would have been 
the result of later linguistic developments during the third millennium, a period 
when highly mobile Yamnaya groups were moving into the Balkans but were 
still very much part of the chain of mutually related Pontic-Caspian — west 
Siberian steppe tribes. Alternatively, some might prefer to place the Greeks in 
western Anatolia by the third millennium bc, from whence they subsequently 
entered Greece. Either way, this puts Indo-European peoples in the Balkans, 
the putative staging area for each of the models of Greek migrations discussed in 
Chapter Three, as well as the appropriate area for the later development of 
either local (Thracian) or transported (Phrygian, Armenian) populations. This 
seems a plausible account for what might have happened but, again matters are 
never so simple. 

Any interpretation of the archaeological puzzle of Southeast Europe during 
this transition period from 4000 to 3000 BC is likely to suffer from a lack of well- 
documented evidence. At the same time, one cannot ignore that a substantial 
number of leading East European archaeologists acknowledge an expansion 
from the Pontic-Caspian to be a major factor in effecting this transformation of 
society and, in general, that they identify these intruders as the earliest Indo- 
Europeans. It is difficult to deny that there is evidence for intrusions unless 
one's preoccupation with explaining all cultural change purely through local 
cultural processess precludes this concept. What remains for us here is to 
inquire whether this appeal to intrusion actually explains the presence of Indo- 
European languages in Southeast Europe and Anatolia. 

In numerous works, Marija Gimbutas has argued that intrusions occurred in 
three waves. Southeast Europe is envisaged as initially having an essentially 
settled, agricultural, matrilineal, peaceful population, with an extremely rich 
artistic and religious vocabulary expressed through ceramics and figurines. 
This society was either destroyed or subjugated by essentially pastoral, warlike, 



patrilineal Kurgan tribes from the Pontic-Caspian steppe. Such a brusquely 
simplified presentation of this invasion hypothesis naturally rests most 
uncomfortably with many archaeologists, and one must surely question an 
explanation that focuses so intensely on elements such as the martial proclivities 
of the steppe intruders. Other factors were surely at play. 

After two millennia of agricultural settlement one might expect there to have 
been a substantial growth in population in Southeastern Europe. Studies which 
have been made of settlement density appear to indicate that the distance 
between settlements, at least in some territories, was clearly diminishing as 
population increased. In Romania, for example, we find Early Neolithic Linear 
Ware sites spaced at intervals averaging 23 to 24 kilometres, while later Pre- 
Cucuteni sites are only 6 to 7 kilometres apart, and during the Cucuteni period 
itself they are only 3 to 4 kilometres apart. Defensive architecture such as 
palisades, ditches and earthen banks appear by the fifth millennium bc, for 
example, at Polyanitsa in Bulgaria, and precede any evidence for steppe 
intrusion, no matter how generously interpreted. Indeed, the expansion of the 
Tripolye culture into such arid areas as the northwest fringe of the Black Sea 
might be explained as the result of population pressure. Quite probably there 
were internal forces within society itself that were setting it on a new course. 

External factors may also be taken into consideration. There is at least some 
evidence for increasing aridity during the Eneolithic which might have 
stimulated more interest in pastoralism. Many of the putatively late features of 
the Eneolithic economy, such as the plough, wheeled vehicles, domestic horse, 
yoke, dairy products, and woollen textiles, emerge in Southeastern Europe 
about the fourth millennium. As Andrew Sherratt has suggested, many of these 
features - such as plough agriculture and increased stockbreeding - would tend 
to enhance the male role in the productive economy. This, in turn, would effect 
a series of social changes which would influence inheritance systems and might 
promote patrilineality, clan associations, warfare, and a diminution in the 
status of the female in society, craft production and, possibly, religious ritual 
such as that so often expressed in Southeast European figurines. All of this 
might suggest that local processes may, in fact, have stimulated the creation of a 
society that matches our expectation of the Indo-Europeans. Nevertheless, 
when the steppe cultures both reflect the Proto-Indo-European vocabulary for 
these economic, technological and social changes, and offer substantial 
evidence for the introduction of the domestic horse, larger woolly sheep as well 
as, possibly, wheeled vehicles into the Balkans, we can hardly dismiss their 
presence as merely coincidental. 

Here we should briefly consider Colin Renfrew's attempt to reverse the 
thrust of these arguments by suggesting that steppe pastoralism was ultimately 
derived from the Balkan cultures. Renfrew rejects a steppe homeland for the 
Indo-Europeans on three grounds: that it is founded on an uncritical reliance 
on linguistic palaeontology (which produces results incompatible with his own 
solution?); that it involves a 'migrationist' model (heretical to the 'immobilist' 
dogma?); and that it does not attend to the processes which led to pastoralism. 



These processes indicate that steppe pastoralism must have developed from a 
mixed farming economy, and hence the Kurgan pastoralists must have been a 
secondary development from primary farming communities. According to 
Renfrew, these farming communities were the Tripolye culture. To the best of 
my knowledge, the suggestion that steppe pastoralism was preceded by more 
stable farming economies has not been disputed by anyone; that these 
communities should have been the Tripolye culture, however, is totally 
contradicted by all of the archaeological evidence so far reviewed. Neolithic 
farming populations spanned the entire region from the Dnieper to the Ural 
before the Tripolye culture had ever expanded towards the steppe, and it was 
east of the Dnieper that we see societies acquiring the necessary prerequisites to 
master the steppe. When some of the regional variants of the Tripolye culture 
did begin to advance into the steppe, about 3500 bc, they were the last, not the 
first to do so. 

If we accept the arguments for some form of expansion from the steppe into 
Southeastern Europe as sufficient evidence for a Pontic-Caspian homeland, 
then we must envision the resulting society of the late fourth millennium B c as a 
hybrid production of native Balkan populations and intrusive Indo-Europeans. 
We can only hypothesize that the changes, both internally and externally 
induced, may have been sufficient to so disturb native development that less- 
numerous intrusive populations, with a more mobile economy, found 
themselves socially more competent to sustain their languages and achieve the 
gradual linguistic assimilation of the indigenous population. Towards the end 
of this chapter we will discuss the processes that may have assisted the 
expansion of the Indo-European languages. 

The evidence suggests that a Pontic-Caspian origin for the Indo-Europeans 
of Southeast Europe and western Anatolia can at least be seriously entertained. 
Here we are hardly talking of proof but rather of a working hypothesis whose 
own integrity is supported by the substantial evidence for actual intrusions from 
the steppe seen in the spread of kurgan burials, the simultaneity of appearance 
of a series of economic traits that are all lexically attested in Proto-Indo- 
European, and the widely accepted interpretation of a structural reordering of 
Southeast European society coincidental with the appearance of these steppe 
intruders. Appeals to authority, naturally, only help underwrite the seriousness 
with which the hypothesis should be considered, not its validity. 

Central and Northern Europe 

Probably the most controversial, and certainly the most discussed, region of 
Indo-European expansion is the territory of Northern and Central Europe, 
from whence we commonly derive the linguistic ancestors of the Celts, 
Germans, Baits, Slavs and possibly others. During the Eneolithic there emerge 
within this region two cultures which exhibit striking cultural similarities 
over extremely broad areas. The slightly earlier and territorially more confined 
Globular Amphora culture appears in the centuries around 3500 BC, and is 



encountered from central Germany to northern iMoldavia. Beginning slightly 
later is the Corded Ware horizon, with its over twenty regional variants. 
These stretch from the Netherlands on the west to the upper Volga and middle 
Dnieper on the east, and are found as far north as Scandinavia and as far south 
as Switzerland. 

The Corded Ware culture has always occupied a position of special 
prominence w ithin the remit of Indo-European origins. Although the evidence 
for Corded Ware settlements is still quite sparse, it has provided a large body of 
mortuary data as attractive to archaeologists on the trail of the Indo-Europeans 
as that recovered from the Pontic-Caspian. Corded Ware burials occur both in 
flat-grave cemeteries and under tumuli. The dead were buried singly or, more 
rarely, in pairs or larger groups, generally in the flexed position on their sides or 
backs. The orientation and burial posture were normally determined by sex: 
males were often buried on their right side, their heads to the west and faces to 
the south; females were buried on their left side and as their heads were oriented 
to the east they too faced south. In this light, the sexual dimorphism in Corded 
Ware burials exhibits what many would assume to be an overt expression of 
Indo-European ideology (we have already seen the same distinction at Nalchik, 
Tulkhar and in the Tazabagyab culture). Some would also regard as significant 
that the dead of both sexes faced south, the direction which the Indo- 
Europeans held to be auspicious as the related words for right, skilful and south 
indicate. It should be emphasized that the mechanics of expressing the 
opposition of the sexes need not have been executed in a uniform fashion among 
all of those embracing the same ideology. Hence, in some variants of the Corded 
Ware horizon males are found on their left side w hile females are buried on their 
right. This need not signal, as has been suggested, the retention of earlier 
'matriarchal' ideologies, but is far more easily explained as a variation in the 
technique of expressing the same ideology, that is, males are buried right side 
up rather than right side down. 

Sexual dimorphism is also exhibited in the nature of some of the grave goods. 
While cord-decorated amphorae may be found in both male and female graves, 
the cord-decorated beakers tend to correlate with males in a number of 

'3$ Typical Corded Ware 
burial of a male accompanied 
with an axe, beaker and 
amphora. The burial is from 
the Corded Ware cemetery of 
Vikletice in Czechoslovakia. 

2 44 


cemeteries. Both of these are ceramic type fossils of the culture that provide it 
with its most common name (Schnurkeramik 'corded ware'). In addition, 
another major trait of the burials is the frequent inclusion of a stone-perforated 
battle-axe with male burials which has provided some of its variants with the 
alternative name of Battle-axe culture (Streitaxtkultur). Other grave goods 
include pottery, arrowheads, bone needles, boar-tooth pendants for males, 
canine-tooth ornaments for females, and copper ornaments. 

A number of factors account for the prominence of the Corded Ware culture 
in our discussion. The paucity of settlement sites traditionally predisposed 
archaeologists to view the Corded Ware culture as predominantly that of a 
mobile, pastoral people. This was reinforced by the presence of tumulus 
burials, a major characteristic of other pastoral peoples, most notably in the 
Pontic-Caspian. The extremely wide territory of Corded Ware occupation is 
suggestive of an expansion across much of Europe which was far more easily 
explained by tying it to mobile pastoralists rather than to settled 
agriculturalists. The presence of weapons in the burials, and the clear sexual 
dimorphism, reinforced the image of an essentially warlike, pastoral people. 
When this was coupled with evidence in the Corded Ware culture for domestic 
horse and wheeled vehicles, there was no problem in completing the image of a 
people which superimposed both their culture and language over a vast 
territory. Both the archaeological evidence, and expectation, encourage us to 
regard the Corded Ware culture as a classic expression of prehistoric Indo- 
Europeans, much as the Eneolithic cultures of the Pontic-Caspian have done. 

The preponderance of what many have considered to be obvious Indo- 
European traits assisted the Corded Ware culture in being identified as not 
merely an Indo-European culture but as the homeland culture of the Proto- 
Indo-Europeans, especially among some German archaeologists of the early 
twentieth century. But, despite strenuous attempts, no one has ever been able to 
generate a serious case for relating the Corded Ware culture to the Indo- 
Europeans of the Balkans, Greece or Anatolia, much less to those of Asia. It is 
also generally admitted that the Eneolithic cultures of the Pontic-Caspian 
cannot be in any way directly derived from the Corded Ware culture or its 
predecessors. It is enough to remember that the Corded Ware culture is 
essentially a third millennium phenomenon and clearly post-dates the earliest 
appearance of Pontic-Caspian Eneolithic cultures. It is for these reasons that 
any attempt to assign an Indo-European identity to the Corded Ware peoples 
must be confined to a state of linguistic development that is ancestral to a 
number of specific Indo-European linguistic groups, but not to all of the Indo- 
European languages. It is also obvious, then, that establishing the correct 
relationship between the Corded Ware and Pontic-Caspian regions is essential 
to the entire problem of determining the earlier territorial boundaries of the 

We begin our review of Corded Ware-Pontic relationships with the 
evidence for chronology. There is an extensive series of radiocarbon dates for 
the Corded Ware horizon which are distributed quite unequally over their 



various regional variants. They indicate a span for the horizon from about 3200 
to 2300 bc, with most dates falling in the third millennium BC. We have already 
seen that the seventy, or so, dates from the Yamnaya culture fall within the same 
very broad range, although they begin somewhat earlier (about 3600 bc), which 
is further emphasized by the fact that no typologically early Yamnaya burials 
have been dated. Although the Yamnaya culture may have begun earlier than 
the Corded Ware, there is no real case for an expansion of Yamnaya invaders 
across the North European plain, producing the Corded Ware horizon. 
Intrusive steppe burials as we previously encountered in Southeast Europe are 
generally absent from the Corded Ware region, and on what little anthropologi- 
cal data we possess, there is no reason whatsoever to associate the Corded Ware 
populations, themselves quite heterogeneous, with the physical type which we 
encounter on the Pontic-Caspian steppe. All of this indicates that any attempt 
to relate the two territories takes a different form from that which we employed 
in the Balkans. 

Generic similarities offer the principal grounds for relating Corded Ware 
with the steppe. Tumuli, single burial, the supine flexed position in some 
instances, the use of especially constructed chambers, inclusion of cord- 
ornamented pottery, tooth pendants, weapons and animal remains, are all part 
of the burial rite of both regions. In a recent comparison of the burial practices 
of the two cultures, Lothar Kilian isolated twenty-three diagnostic features. He 
argued that the Corded Ware burials possessed a series of traits not found in the 
Pontic-Caspian - amphorae, cord-decorated beakers, battle-axes - which are 
the essential markers of the Corded Ware culture. In contrast, the steppe 
burials utilized egg-shaped pottery, hammer-head pins, ochre and a variety of 

139 Chronology of Corded Ware and the steppe cultures (number of dates in each sample 
indicated in parentheses). 



burial postures unknown in the Corded Ware horizon. While there may be 
some generic similarities, Kilian concluded that the specific differences do not 
support an historical connection between the two regions. For many years 
Alexander Hausler, an authority on the mortuary practice of both the Pontic- 
Caspian and the Corded W are regions, has perennially campaigned against any 
possible association between them. He especially emphasizes that the ritual 
core of the Corded Ware burials involved depositing the deceased in the flexed 
position on his or her side, depending on sex. This signalling of the sexual 
dichotomy through burial posture is absent from the steppe region and, he 
argues, denies it any role in the formation of the Corded Ware horizon. He 
further emphasizes that the Corded Ware burial rite can be derived from local 
Central and North European cultures without recourse to steppe intrusions. 

Objection to a Pontic origin for the Corded Ware horizon is not limited to the 
evidence for burials but extends throughout the range of arguments. Evzen 
Neustupny, for example, dismisses the entire concept of pastoral Corded Ware 
tribes as being incompatible with both the environment of Northern Europe 
and the archaeological evidence for agriculture in the Corded Ware culture. 
This, he argues, suggests a local evolution of the Corded Ware horizon from its 
earlier Neolithic and Eneolithic populations. 

It should be emphasized that there is no widely accepted solution to the 
problem of Corded Ware origins. Ivan Artemenko recognizes the existence of at 
least four major camps: those who support a western origin between the Rhine 
and Vistula (such as Ulrich Fischer, Alexander Hausler and Karl Jazdzewski); a 
Vistula-Dnieper origin (Dmitry Kraynov, Raisa Denisova and Miroslav 
Buchvaldek); an origin in the forest-steppe zone of the Middle Dnieper region 

140 Corded Ware groups in 
Eastern Europe. 



(Ivan Artemenko, I. K. Sveshnikov, V. P. Tretyakov, Sofia Berezanskaya, 
N. Bondar, and others); and a steppe origin (Gustav Rosenberg, P. V. Glob, 
Karl Struve, Marija Gimbutas, Aleksandr Bryusov and Valentin Danilenko). 
Anticipating the serious implications of granting to the Corded Ware culture an 
essentially local origin, we are impelled to survey all the possibilities that might 
indicate an historical link with the steppe. 

We turn first to the Middle Dnieper culture, one of the local variants of the 
Corded Ware horizon, since it occupies the frontier between the Pontic and the 
rest of the Corded Ware horizon. The contact zone falls on the Middle Dnieper 
itself. Ivan Artemenko divides the Middle Dnieper culture, known from more 
than 200 sites, into three phases, the earliest of which is found in a number of 
graves located southwest of Kiev. As a rule, these burials were inserted into the 
kurgans of the local variant of the Yamnaya culture that had extended into the 
forest-steppe. They were accompanied by cord-ornamented pottery and 
perforated battle-axes. Artemenko has suggested that the origin of these burials 
is to be sought in the Sredny Stog, Yamnaya and Late Tripolye cultures. This 
phase is then followed by a middle period where the classic Corded Ware 
amphorae and beakers appear. Artemenko, and others, see in the Middle 
Dnieper culture the roots of some of the other local variants of the Corded Ware 
horizon in easternmost Europe, but no one seriously entertains the hypothesis 
that the Middle Dnieper culture is the earliest expression of the Corded Ware 
horizon, the one from which it later expanded westwards. 

While the Corded Ware variant most proximate to the steppe cannot be seen 
as the point of origin for the Corded Ware culture, many archaeologists still 
envisage a major influential role here from the steppe cultures. One of the 
foremost authorities on the Corded Ware horizon, Miroslav Buchvaldek, has, 
on several recent occasions, addressed the problem of Corded Ware origins. He 
emphasizes the methodological problems of demonstrating whether the 
Corded Ware horizon was the product of an invasion or local development. 
According to both models, the number of sites belonging to either the intrusive 
culture or to the transitional phase between a former Neolithic culture and the 
subsequent Corded Ware horizon would be extremely small, a generation's 
worth of sites perhaps constituting no more than 5 per cent of the potential 
archaeological remains. The odds of discovering enough intrusive or transi- 
tional sites to establish a convincing pattern would consequently be extremely 

Buchvaldek concentrates on the origins of the 'A' or 'common European' 
horizon of Corded Ware finds which embrace a series of axes, beakers and 
amphorae displaying remarkable similarities and which may be found from the 
Netherlands to the Ukraine. Those who support an indigenous development 
for this horizon make recourse to 'influences' spreading across Europe; 
Buchvaldek regards these as insufficient to explain the uniformity of the 
cultural phenomenon against the great variety of Europe's background 
cultures. The problem with an invasion, however, is also admitted - it requires 
a source, and there simply is not one that is particularly convincing. However, 



Buchvaldek does note the similarities between the Yamnaya and Corded Ware 
cultures, such as tumulus, flexed supine burials, corded ornament, heavily 
pastoral component, absence of settlements, and several other items of grave 
accompaniment. He thus considers that the Corded Ware culture shows 
stronger ties with the Yamnaya culture than it exhibits with any other culture, 
though he admits that there are also striking differences between these two 
cultures. Also, parallels exist between the Corded Ware and the Tripolye, 
TRB, and Globular Amphora cultures. Buchvaldek opts for a migration or 
infiltration of small human groups from some region where all of these cultural 
influences or components may have come together. The actual location is 
uncertain and Buchvaldek expressly admits that it is on the basis of both 
'intuition and hope' that he believes a home for the 'A' horizon may be found in 
the poorly known region between the Vistula and the Dnieper. 

As the most superficial glance at a map will indicate, the Corded Ware 
horizon occupies essentially the same general territory as the preceding TRB 
culture. Where it exceeds the borders of the TRB, such as the Fatyanovo 
culture of the upper Volga, or the Baltic Haff culture, this is clearly a later 
intrusion outside the core area. For this reason, it is impossible to consider 
Corded Ware origins without taking into account the preceding TRB culture. 
This was by no means a uniform cultural entity that can easily be summarized 
with a few general statements. Burials, for example, ranged from simple, 
earthen flat graves to long barrows and megalithic tombs. Although this range 
was wide it can be argued that it offered a poor ancestry for the tumulus burials 
often encountered in the Corded Ware horizon; one variant of the TRB culture, 
however, proves the exception. 

The Baalberge group is at present known from approximately 200 graves 
distributed in central Germany east to Bohemia. The ceramics accompanying 
the burials include characteristic TRB forms plus some vessels indicating more 

Baalberge Culture 

Area of Steppe 


141 Baalberge and the steppe cultures. 





easterly influences, such as the Bodrogkeresztur and Baden cultures. What is 
most important is that the Baalberge group often buried their dead in stone cists 
over which an earthen tumulus was erected. Those arguing for a local Corded 
Ware origin could point to the Baalberge group which provides a suitable 
ancestor for the Corded Ware grave structure plus distant ceramic prototypes 
for Corded Ware beakers and amphorae. On the other hand, some have argued 
that the Baalberge group was the result of steppe pastoralists who had 
proceeded along the Danube into central Germany where they effected the new 
burial rite. The difficulty in pursuing the latter model is that the nearest 
positively identified remains of steppe-like kurgans lie to the east of the Tisza 
river some 500 kilometres short of the territory in which the Baalberge group 
emerges. Unlike the Siberian steppe, this area is not a terra incognita within 
which we can optimistically hope for the discovery of classic steppe kurgan 
burials. Also, physical anthropology from the Baalberge graves offers no 
grounds for associating them with the more easterly Pontic intruders. 

Another door we may try concerns the origins of the Globular Amphora 
culture. In her more recent publications on the expansions of the Pontic- 
Caspian tribes, Marija Gimbutas has regularly indicated that she believes the 
Globular Amphora culture to have been a major component in the creation of 
the Corded Ware horizon, a theory we encountered with Miroslav Buchvaldek. 

Like the Corded Ware horizon, the Globular Amphora cultures emerged out 
of a TRB substrate. However, Gimbutas argues that upon this base was 
superimposed a ruling stratum from the Pontic region. The evidence, she 
argues, is to be found in the complete congruency of the burial rites of the two 
areas - stone cists, cromlechs, stelae, ritual burial of animals including horses, 
and the practice of suttee, executing the wife and possibly other members of the 
family on the death of the husband. To this we could add the occasional use of 
ochre and kurgans in the eastern region of the Globular Amphora culture. She 
compares the typical vessel of the culture, the globular pot with two to four 
small handles below the neck, with those from the Lower Mikhavlovka group. 
The Globular Amphora culture appears to have involved mixed agriculture; 
remains of cattle, sheep, pig and horse have been found on sites. 

142 Generalized distribution 
of TRB (dotted line} and 
(r lobular Amphora (shaded 
area) cultures. 



a smaller compartment. 

Cartographically, there are difficulties in trying to sustain a steppe origin for 
the Globular Amphora culture, at least one that is confined to the comparisons 
cited by Gimbutas such as the stone cists and stelae of the Lower Mikhaylovka 
and more southerly steppe cultures. Again, we are speaking of a distance in the 
order of 500 kilometres between these two cultures. Moreover, Hauslcr 
observes that there are only two stelae in the Globular Amphora culture and 
that these are from northern Poland, while the ritual burial of animals, also cited 
by Gimbutas as evidence for steppe influence, was known already from the local 
TRB culture. Certainly such authorities as the Polish archaeologist Tadeusz 
Wislanski have generally opted for a Globular Amphora origin in the basins of 
the middle and lower Oder and Warta. In a recent review of Globular Amphora 
remains in the Ukraine, I. K. Sveshnikov favours a movement into the Ukraine 
from the west and does not even entertain the hypothesis of a steppe origin. 

In order to attempt yet another link between Northern Europe and the 
steppe we need to recede still further in time. Alexander Kosko, for example, 
has recently argued that the steppe tribes, through both pressure and 
assimilation of their Tripolye neighbours, destroyed the traditional Dniester- 
Bug frontier that divided the North European cultures from the Pontic- 
Caspian. This is evidenced by the appearance of the Matw y group of the TRB 
culture, in the Vistula region, which shares a series of ceramic similarities with 
the later Tripolye culture. In short, Kosko is arguing for the creation of an 
interaction sphere that connected the Pontic with Northern Europe during the 
TRB culture, somewhat along the lines we have suggested between the Pontic 
and the Balkans-west Anatolia. The evidence for this, however, still seems to be 
extremely sparse. 

Jan Lichardus has suggested that the steppe influences are to be found in the 
formation of the TRB culture itself. The earlier phase of the TRB culture, the 
B-phase, is linked to cultural impulses or immigrations from the Danubian 
region northwards, especially the Rossen culture (others would argue that the 



Lengyel culture was the major donor). TRB ceramics, polished stone axes, 
burials in flexed position in flat graves are all products from the south. The 
slightly later TRB A horizon, however, he believes to be underivable from 
both the Danubian region and the local Ertebolle culture of Northern Europe. 
This horizon sees the introduction of extended burial, occasional use of ochre, 
knobbed hammer-axes, and several other features which Lichardus believes 
can only be derived from the east. 

Lichardus resurrects an old theory, first suggested by the Danish archaeo- 
logist C. J. Becker, that the origin of the TRB culture may lie in the Ukraine. 
While the earlier theory has long been abandoned, Lichardus suggests that 
there is now sufficient evidence to revive it, since the Dnieper-Donets and 
Sredny Stog cultures do provide the most convenient source for what is new in 
TRB A. Here we find extended burials, use of ochre, similar orientation of 
burials and prototypes of the TRB A axes, as well as the early use of corded 
ornament on pottery. Wheeled vehicles are known both on the steppe and in the 
TRB culture. Of particular interest is the recent identification by Lichardus of 
a pair of late TRB antler psalia which are similar to those recovered in the 
Sredny Stog culture and elsewhere on the steppe. Lichardus has undertaken 
some experimental archaeology and has demonstrated that such devices serve 
admirably as cheekpieces for controlling horses. The evidence for domestic 
horse and horse-riding might, then, also link the two regions. The problem of 
route is resolved, according to Lichardus, by the fact that Dnieper-Donets sites 
are known to extend as far to the northwest as the Pripet. From there, mobile 
pastoralists, with horses and wagons, could push west to the Vistula and 
throughout Northern Europe. As TRB A is one of the major components of the 
Baalberge culture, we thus obtain a steppe presence throughout Northern and 
Central Europe that can also be ancestral to the Globular Amphora and Corded 
Ware horizons. 

Opponents to Lichardus's theory assert that his model for an external origin 
for these TRB traits is unnecessary, unsupported by specific archaeological 
evidence and inherently unlikely. Brigitte Hulthen, for example, notes that we 
may derive the TRB ceramics from the previous local Ertebolle culture. 
Alexander Hausler argues that there was no difference between TRB B and 
TRB A burial rites, and that the extended position is precisely what one might 
have expected from a Neolithic culture whose substrate was the hunter-fishers 
of Northern Europe across which extended burial was the norm. The 
similarities between the Dnieper-Donets and TRB burials are largely generic to 
all of the sub-Neolithic populations of Northern Europe, from Scandinavia to 
the Urals. Furthermore, it is difficult to appeal to the cultural practices of two 
different cultures - Dnieper-Donets and Sredny Stog - in order to explain the 
intrusive element in the TRB culture. Sredny Stog sites are in no way 
proximate to the area in which the TRB culture appears, and the Dnieper- 
Donets sites along the Pripet have not been shown to provide suitable parallels 
for the TRB A horizon. Essentially, a number of German reviewers of 
Lichardus's work wonder whether an appeal to a migration of horse-mounted 



pastoralists from the steppe is either necessary or even plausible for Northern 
Europe. Moreover, Soviet archaeologists generally view the presence of the 
TRB culture in their own territory as an intrusion from the west rather than the 
reverse. Finally, Dmitry Telegin, the leading authority on both the Dnieper- 
Donets and Sredny Stog cultures, dismisses the thesis of a TRB origin from the 
Dnieper-Donets culture. 

Lichardus's model is a variant of a broader explanation of the cultural change 
seen throughout both Northern and Central Europe in the Late Neolithic. 
Marija Gimbutas does not confine the effects of her 'first wave' of steppe 
pastoralists to Southeastern Europe. She argues that these intruders not only 
pushed local Balkan cultures westwards but also that they themselves passed 
further west along the Danube to stimulate the formation of 'kurganized' Late 
Neolithic cultures. Among these cultures Gimbutas lists the later Lengyel and 
Rossen cultures of Central Europe. This process of hybridization is set to the 
period about 4500 to 4000 bc. The evidence for 'kurganization', at least that 
published so far, would appear to be a series of very general cultural changes: 
reliance on stockbreeding rather than on agriculture; preference for defensive 
locations of sites; small settlements; solar symbolism on pottery; and the 
appearance of the horse. If we read between the Jines, this case seems to suggest 
that the transformations seen across later Neolithic and Eneolithic communi- 
ties in Central Europe were in the general direction of increasingly pastoral and 
aggressive tendencies in society, which could best be explained by the intrusion 
of pastoralists who took their ultimate origin from the steppe. Hence, defended 
sites from Cernavoda I in Romania west to the Rossen culture are all linked 
together as 'kurganized' cultures involving local populations and an intrusive, 
normally dominant, steppe element. If one were to accept this model, then we 
would be provided with steppe-based ancestors to the TRB culture as a whole 
(which, curiously enough, Gimbutas generally regards as the antithesis of the 
steppe cultures). The level of continuity that one encounters between the 
Linear Ware culture and the Rossen culture in, for example, settlement types 
and long houses, appears to be largely ignored, and where the argument 
concerns a steppe element, it seems mainly confined to very broad categories of 
cultural change, regularly regarded as the product of local processes rather than 
of foreign intruders. It is not that the model is wrong; there is just not enough 
evidence proposed to evaluate it seriously. 

If we adopt the broad perspective and consider in general the formation of 
the large, cultural entities such as the Globular Amphora and Corded Ware 
horizons, we encounter the perennial problem of how such cultural entities 
might have come into existence. Invariably there are those who would support 
autochthonous development, with cultural convergence or 'peer-polity interac- 
tions' over a broad area to account for widescale cultural similarities. Others 
will generally opt for migrations that carry a similar culture over diverse 
substrates. Few would be so doctrinaire as to adhere only to the stadial 
development of local groups who just happen to share broad categories of 
ceramic types, decoration, weapons and burial rites over vast areas. Certainly 



the intrusive nature of Corded Ware in Upper Lusatia, the Baltic, the upper 
Volga and possibly Holland all require some external impulse if not actual 
population movements. On the other hand, even when one excludes these 
supposedly later peripheral areas, and pares down the size of Corded Ware 
distribution, it is still impossible to select confidently an area of origin. 

We have now examined most proposals that attempt to give the steppe 
cultures a role in the cultural transformations of Central and Northern Europe. 
We have seen how they generally, if not universally, fail to present the type of 
primary evidence for migration that we have seen in the Balkans. On the other 
hand, much of the criticism levied at the steppe camp has concerned their 
failure to produce classic Pontic-Caspian burials west or north of the Tisza, and 
one must wonder whether this is necessary. The distance involved between the 
steppe and the putati vely kurgan-dominated cultures of Central Europe is quite 
substantial, and the model for steppe intrusions does not require a single event- 
horizon of pristine horse-riding warriors ravaging the banks of the Danube. 
Progressive change as one moves from the steppe towards Central Europe is to 
be expected, and the failure of the Baalberge burials to mirror precisely the 
steppe kurgans of the Ukraine is not totally devastating to the theory of 
migrations. Similarly, aspects such as the sexual dichotomy in burial practice, 
advanced by Alexander Hausler as the 'ritual core' of the Corded Ware culture 
and hence a primary reason to exclude a steppe ancestry, is in turn dismissed by 
Nikolai Merpert as trivial compared with the more fundamental similarities 
between the two traditions. 

Nevertheless, the archaeological evidence advanced for the origins of the 
Corded Ware horizon has, so far, failed to make a thoroughly convincing case 
for population movements or intrusions, the minimum requirement of our 
search for the trajectory of the earliest Indo-Europeans. In our evaluation of the 
archaeological evidence we are tied procedurally to assume local development 
unless demonstrated otherwise, and the case for intrusion simply is not strong 
enough. There is no question that a number of notable and quite knowledgeable 
archaeologists support the concept of some form of genetic relationship 
between the steppe and the Corded Ware horizon - either by simple intrusion 
or by some complex process of assimilation and convergence within the forest- 
steppe zone between the Dnieper and Vistula. They provide us with enough 
evidence to charge, but not enough to convict. 

What, then, are the consequences of rejecting a Pontic presence in Central 
and Northern Europe during the Eneolithic? Both Lothar Kilian and 
Alexander Hausler have followed their rejection of Kurgan intrusions to its 
logical conclusion. If the Pontic-Caspian cultures cannot be derived from 
Northern and Central Europe, nor the latter from the former, then they must all 
be included within the Indo-European homeland. Since they admit of no 
Neolithic or Eneolithic relationship between the two major regions they must 
assume that it was during the Mesolithic or Palaeolithic that a vast linguistic 
continuum existed that connected the North Sea with the Volga-Ural. Out of 
this continuum there later evolved the Indo-European languages. If Proto- 



14$ The Indo- 
European homeland 

according t0 
Lothar Kilian. 

Indo-European is defined in the strict sense to the period 4500-2500 bc, then 
TRB, Globular Amphora, Corded Ware, Sredny Stog, Yamnaya and the other 
steppe cultures should all be grouped together as Proto-Indo-European. This 
theory has wide currency though it is normally more implicitly maintained 
than explicitly developed. Many Soviet archaeologists also assign the Pontic- 
Caspian solely to the ancestors of the Indo-Iranian languages and assume that 
the other European languages, especially Baltic and Slavic, developed 
northwest of the Pontic from local cultures. 

The models proposed by Kilian and Hausler seem to require a linguistic 
continuum some 2,000 to 3,000 kilometres long, during the Mesolithic or 
Palaeolithic. They postulate no archaeological phenomenon, not even a major 
lithic techno-complex, which might suggest that there was some form of 
historical interrelationship across this continuum. Even archaeologists such as 
Janusz and Stefan Koslowski, who frequently paint in the Palaeolithic and 
Mesolithic cultures of Europe with an exceedingly broad brush, do not propose 
some form of cultural entity that embraced this entire region. While it is true 
that an archaeological entity need not correlate w ith a linguistic one, how can 
one snatch at diverse archaeological cultures and merely assume that they all 
spoke related languages? We shall have to make a go at both evaluating this 
theory and defending it against our own objections. 

First, it seems highly likely that an Indo-European homeland extending 
from the Rhine to the Dnieper or the Volga has certain problems of plausibility. 
Separated for thousands if not tens of thousands of years, the various bands of 
hunter-fishers distributed across this continuum would surely have evolved 
vastly different languages in terms of phonetics, grammar and vocabulary no 
matter what their original genetic relationship. Yet the model of such a 
homeland requires us to believe that they maintained an essentially unchanged 
phonetic or grammatical system up until about the fourth or third millennium 
bc, despite the fact that intense mutual interrelationships across this region are 
specifically denied by those who oppose a later Pontic-Caspian influence in 



Northern and Central Europe. We are asked to believe that the ancestors of the 
Tocharians or Indians on the extreme east of this continuum shared the same 
vocabulary (including all those items that could not have entered the 
vocabulary prior to the Neolithic-Eneolithic), phonetic system and grammati- 
cal structure with the ancestors of the Germans and Celts who would have been 
situated somewhere on the western end of this continuum. All of this seems 
unlikely, and we need to emphasize again that linguists generally regard genetic 
relationships between languages spoken as long ago as the Mesolithic or 
Palaeolithic to be irretrievable because of the magnitude of linguistic change 
over many thousands of years. The result of expanding the Proto-Indo- 
Europeans across Mesolithic or Palaeolithic Europe sets the clock on linguistic 
change far earlier than is plausible. 

One solution to our problem would be to abbreviate the size of this earlier 
territory and to seek some demonstrable connection between at least the Vistula 
and the Pontic-Caspian prior to the period we have argued for Kurgan 
expansions. Here we can return to Dmitry Telegin's proposal that the Dnieper- 
Donets culture can be included among a broad group of Vistula-Dnieper sub- 
Neolithic cultures. These would include the Narva, Valdai and Comb-pricked 
Ware cultures of Poland. It should be emphasized that Telegin himself does not 
associate these cultures with the Proto-Indo-Europeans (but rather with the 
ancestors of the Baits and Slavs), yet an historical association across the Vistula 
to the Dnieper during the Neolithic would provide an historical bridge between 
the two regions. With no specific archaeological justification this solution has 
been grasped, perhaps in desperation, by others who have also floundered on 
the uncertainty of the origins of the Corded Ware horizon. With cemeteries 
arguably related to the Dnieper-Donets culture appearing on the middle Volga, 
such as at Sezzhee, we might then argue for an expansion of this continuum 
from the Vistula to the Volga by the fifth millennium bc. This requires us to see 
the Dnieper-Donets culture as the dominant partner in the creation of the 
Pontic-Caspian steppe communities and its northwestern cousins as the 
primary substrate in the creation of the Globular Amphora and Corded Ware 

Objections to this solution are several. Most, if not all, archaeologists 
engaged in the study of the Pontic-Caspian accept the hypothesis that the 
Dnieper-Donets culture was swallowed up, both in terms of culture and in 
terms of physical type, by other populations from the steppe. Although Telegin 
describes the Dnieper-Donets as a physical wedge driven from the north 
towards the steppe, he also indicates that it was assimilated by the Sredny Stog 
and Yamnaya cultures. Gimbutas specifically dismisses it as a local substrate 
assimilated by Indo-Europeans from the Volga-Ural region, a general course of 
development also accepted by other archaeologists. Moreover, we have the 
objections of Alexander Formozov who denies particularly close genetic 
relations with the northwest because Dnieper-Donets sites only expand in this 
direction in its latest phases. Similarly, V.P. Tretyakov emphasizes the 
independence of the Upper Dnieper Neolithic culture from the Dnieper- 



Donets. Most importantly, the other cultures of Poland and the Baltic are sub- 
Neolithic and do not reveal the culture reconstructed for the Proto-Indo- 
Europeans until they have experienced the expansion of the Corded Ware 
horizon into their territory. 

Another route out of this maze is the traditional fall-back position to the 
Linear Ware culture which does extend from the Rhine to the western 
Ukraine. 37 Its archaeological descendants - Rossen, Stichbandkeramik and 
Lengyel - occupied Central Europe and have often been credited as the external 
stimulus to the Neolithic economy of Northern Europe. Moreover, both the 
Linear Ware and Lengyel cultures penetrated into the northwestern region of 
the Ukraine along the upper tributaries of the Dniester. If the Linear Ware 
culture could be seen to have served as an integral component of Tripolye, we 
might then follow this linguistic vector as far east as the Dnieper. Here, even 
with good will, this theory begins to collapse, since we must also demonstrate 
an historical relationship between the Linear Ware and the steppe region where 
all the archaeological evidence denies any role to western intrusions during the 
Neolithic. Indeed, the entire trajectory of movement is in the opposite 
direction. Moreover, many would adhere to the hypothesis that the Linear 
Ware culture, at least in Central and Eastern Europe, was essentially a 
projection of Balkan populations into the Danube region which would then tie 
the Linear Ware culture to the same linguistic objections we have seen for 
Southeast Europe. It is no easier to derive the steppe cultures from Central 
Europe than it is to derive the North European cultures from the steppe. 

Readers may feel that the author has betrayed them down an endless series of 
cul de sacs. Nevertheless, this is the current state of research into Indo- 
European origins and this seemed the best way to convey why the issue is by no 
means resolved. Ultimately, we have a remarkably unsatisfactory set of choices. 
We can accept a Pontic-Caspian homeland despite the fact that it still appears to 
be archaeologically undemonstrated, even under the most liberal canons of 
proof, in explaining the Indo-Europeans of Northern and Central Europe. 
Alternatively, we might wish to opt for a broader homeland between the Rhine 
and Volga during the Palaeolithic or Mesolithic which resolves the archaeo- 
logical issues by fiat but appears to be linguistically implausible. Perhaps our 
only recourse is to return to our strict definition of the Proto-Indo-European 
homeland as where the Indo-European languages were spoken in the period 
4500-2500 BC. By the end of this period it is reasonable to assume that they were 
spoken from the Rhine to beyond the Ural. How they achieved that position is 
still a problem. 

The Process of expansion 

As Ernst Pulgram observed thirty years ago, there are three ways by which we 
might imagine a language to expand: the migration of complete populations; 
infiltration of an area by small groups; or diffusion. As the last hypothesis has 
never been encountered, we assume that the borders of Proto-Indo-European 



or the individual Indo-European languages spread by migration or intrusion. 
We may suspect that the distinction between migration and intrusion is not 
particularly fundamental. Logistically, it is easier to move small numbers of 
people around rather than large numbers, but the traditional tendency to assure 
one's audience that you intend only small intrusions probably has more to do 
with the embarrassment one associates with long-ridiculed invasion hypoth- 
eses. We are always wary of suggesting models of expansion that will be 
caricaturized as hordes of frenzied Aryans bursting out of the Russian steppe 
and slashing their way into the comparative grammars of historical linguists. 
Similarly, we tend to shy away from making the Indo-Europeans anything but a 
numerical minority in their relations with the substrates which they eventually 
assimilated. On a broad scale this is probably correct, yet we have no reason to 
exclude the development of Indo-European majorities in many local situations 
that brought about the progressive assimilation of non-Indo-European- 
speaking neighbours. With these caveats behind us, we can turn to some 
possible models that might help to explain the incredibly successful expansion 
of the Indo-European languages. 

It is reasonably certain that the process by which the Indo-European 
languages expanded was not uniform but rather dependent on particular 
circumstances. An expansion east of the Urals, for example, may have taken 
Indo-Europeans into steppe regions that were sparsely populated. Possessed 
with a more productive economy which was especially suited to the exploitation 
of the open steppe, the numbers of Indo-European speakers would predictably 
have swelled to become the dominant language(s) across the entire south 
Siberian steppe to the Yenisey. Local hunter-gathers could have been rapidly 
assimilated into the social organization of the Pontic-Caspian intruders. 

When we turn to the spread of the Indo-European languages into areas 
where we find previously settled agricultural populations, we must seek other 
explanations for their expansion. These would include interactions with the 
proto-urban settlements of Central Asia, the substantial villages of the Tripolye 
and other Southeast European cultures. Here, within the context of overall 
regional settlement, it is likely that the natives were numerically superior. To 
explain how the Indo-European languages were ultimately adopted, we must 
return to first principles. 

When two languages come into contact, people speaking one of them do not 
immediately abandon their own and adopt the language of the other. A 
prerequisite to language shift is societal bilingualism. This may remain quite 
stable over a long period but in the case of Indo-European expansions it was 
obviously a prelude to the adoption of Indo-European. We assume for the 
expansion of the Indo-European languages that native populations became 
bilingual for a time, speaking both their own language and adopting that of the 
intruder. Normally, social context determines which language is spoken. For 
example, natives might have spoken their own language at home but Indo- 
European in the market place or at ceremonial activities. If the intrusive 
language is employed in more and more different contexts, it will eventually 



lead to the total replacement, or language death as it is sometimes called, of the 
native language. In Western Europe we have witnessed Latin move from the 
vernacular to the language of the church and the learned classes to that of only 
the church, and now even that role has been greatly diminished. Languages 
upon the point of extinction are normally carried to the grave by the older 
members of the community when the younger members have failed to learn it. 
This process can happen within three generations. An immigrant family in 
Britain, for example, may speak exclusively Hindi while their children become 
bilingual. They in turn decide to raise their children exclusively in English. 
Within three generations grandparent and grandchild can no longer communi- 
cate. This process should have occurred across Eurasia wherever we find the 
expansion of the Indo-European languages. 

It is clear, then, that the crux of the issue is bilingualism and how it was 
induced. Without state coercion, we do not imagine that second languages are 
forced upon people. Rather, bilingualism is induced when the context of speech 
requires the use of the new language if one wishes to obtain better access to 
goods, status, ritual or security. The success of Indo-European expansions 
should have been due to their ability to offer such advantages to the populations 
with whom they came into contact. There are several additional factors which 
may have favoured them in such contact situations. 

Whatever side we view the Pontic-Caspian from, be it from the Volga-Ural 
region or the southwest corner of the Black Sea, we seem to encounter the more 
mobile economies of pastoralists. It is from these steppe regions that we trace 
Indo-European intruders into Asia and the Balkans. In both cases they 
eventually come into contact with stable village communities practising mixed 
agriculture. There is a considerable body of evidence concerning the nature of 
interrelationships between pastoralists and agriculturalists, especially in the 
Near East among both Iranian and Semitic-speaking populations. Pastoral 
economies display a potential for positive growth in that they are dependent on 
the natural productivity of the herd. Good years permit the growth of the herd 
which can be translated into both wealth and status, while bad years, caused by 
climate or disease, can be disastrous. Either way, the pastoral economy offers 
considerable opportunity for social mobility which is much more difficult for 
settled agriculturalists. In his examination of nomad-villager relationships in 
the Near East, Fredrik Barth observes that, in the long term, it is normally the 
pastoralists who are at an advantage in terms of exchange relations and capital 
accumulation, which ultimately leads to their dominance in local systems of 
stratification. In short, there will be a tactical advantage to mobile Indo- 
Europeans in their earliest expansions where they would have assumed 
positions of local dominance. As their mobility was greatly enhanced by the use 
of the horse, expanding their territorial movements over five times that of their 
pedestrian neighbours, they would also have had the opportunity to exploit and 
perhaps dominate larger political territories. 

When we return to the archaeological evidence of Southeast Europe we 
encounter what V. Gordon Childe once termed the 'Late Neolithic Crisis'. He 



saw this reflected in the collapse of village settlement and the adoption of 
increasingly more pastoral modes of production. We have seen how Marija 
Gimbutas has attributed this structural change in society to the intrusion of 
Indo-Europeans whose presence, if not warlike activities, accounted for the 
collapse of the major Eneolithic cultures of Southeast Europe. Others have 
emphasized local processes to account for this collapse. Intensive agricultural 
productivity against a background of population growth might well have taken 
agriculture to its prehistoric limits and have required the adoption of 
pastoralism to maintain growth. Climatic change, especially associated with the 
shift from the Atlantic to the Sub-Boreal, has also long been regarded as an 
important factor. Graeme Barker has observed that the climatic change in the 
Balkans would have resulted in decreasing agricultural productivity, and an 
opening of the landscape, which would have been more suited to pastoralism 
than to agriculture. Certainly, in some pollen profiles from the region, we find a 
decrease in cereal pollen and a reforestation of the landscape contemporary with 
the intrusion of steppe pastoralists. Finally, technological developments 
associated with Andrew Sherratt's 'Secondary Products Revolution' would 
have been another contributory factor. Whatever the causes, we seem to find 
both the abandonment of earlier, stable communities and the reduction of the 
size of settlements with their subsequent dispersal. Jan Makkay, for example, 
notes that, in Hungary, the earlier Neolithic Tisza culture was represented 
by twenty-eight substantial sites. Later, in the Eneolithic, settlement was 
dispersed into about 243 small single-layer Tiszapolgar sites. Petre Roman and 
Sebastian Morintz have also shown how the structural reordering of Eneolithic 
society in Romania resulted in the collapse of former cultures and the 
appearance of scattered settlements on islands or in upland caves. 

The various explanations for the changes in Southeast European Eneolithic 
society are all reasonably persuasive, and they may all have been important 
factors. The fourth millennium BC was a period of social fragmentation and 
apparently increasing pastoralism. Whether they induced these changes or, 
more probably, were coincidental with them, intrusive steppe populations may 
well have been at a social advantage with respect to the native cultures which 
were less easily able to adapt to new conditions. 

In their interactions with the local populations, we should not immediately 
leap to simplistic (and typical) assumptions where superiority in mobility or 
weapons can be directly translated into linguistic dominance. Again, Fredrik 
Barth provides archaeologists with a cogent reminder that we need to look 
beyond apparent military advantages in determining the ultimate formation of 
super- and substrates. Barth examined the linguistic relations between the 
Pathans and Baluchi on the Afghan-Pakistan border. The Pathans were the 
more numerous, the wealthier, better armed, and even possessed a better 
military reputation. Nevertheless, it is the Baluchi who have been making the 
sustained linguistic assimilation of the Pathans. The Baluchi social structure is 
hierarchic and encourages vertical relationships between local leaders and 
clients. The various bands offer opportunities for social advancement within 



these hierarchies, and displaced Pathans in a frontier situation are attracted 
individually and in groups to join Baluchi communities. On the other hand, the 
more egalitarian society of the Pathans was ill-suited to absorb foreigners who 
could only enter it either in roles despised by the Pathans or by undertaking a 
more complicated process to being admitted as an equal in Pathan society. The 
nub of the issue here is not weapons, wealth or population size but the social 
permeability of the competing social organizations. As numerous historical 
instances testify, pastoral societies throughout the Eurasian steppe are typified 
by remarkable abilities to absorb disparate ethno-linguistic groups. Indo- 
European military institutions may have encouraged membership from local 
groups in the form of clientship which offered local populations greater 
advantages and social mobility. It is only possible to speculate about the nature 
of Eneolithic social structures in Southeast Europe. But with the traditional 
centripetal tendencies of settled agricultural societies, they may have been far 
less open to assimilating the mobile pastoralists who lived among them. 

Towards the end of the fourth millennium bc the earlier processes of 
disintegration seem to have been reversed. We find the integration of former 
cultures into new groupings by which the Pontic, Balkans and northwest 
Anatolia appear all to be linked in terms of ceramics, metallurgy and 
architectural forms. This circum-Pontic interaction sphere now sees the 
emergence of citadels which may have served the centralizing tendencies of new 
elites. From our perspective, these could have continued the process of 
spreading the Indo-European languages where dependent populations were 
encouraged to learn the languages of the intruders in order to gain access to the 
products or services of the new elites, whom we may suppose spoke Indo- 
European languages. 

These are only general and quite hypothetical models to suggest how the 
Indo-European languages may have expanded. We may conclude this section 
by laying to rest one fallacy that has often appeared in the past. A tendency to 
see the Indo-European languages as inherently those of the superstrate can be 
found widely in literature on the Indo-Europeans. This form of 'Aryan 
manifest destiny' ultimately calls into question the whole process of expansion. 
In any event, our prehistoric evidence suggests that Indo-Europeans did not 
always maintain their elite position. If they did penetrate the Caucasus, we 
know of no Indo-European language in the Kuro-Araxes region that survived 
into history. Similarly, the region of the Fatyanovo culture of the upper Volga, 
as well as other Corded Ware variants in the east Baltic, clearly succumbed to 
Finno-Ugric speakers. And, from our better-controlled historical evidence, we 
know that Indo-Europeans succumbed to the Hungarians in Europe, and we 
witness the lightning expansion of the Turks, largely at the expense of 
Indo-Europeans. Turkic speakers are probably to be credited with the 
linguistic death of Tocharian as well as with the assimilation of numerous 
Iranian speakers across the Asiatic steppe, and ultimately with the effective 
collapse of Greek as the major language of Anatolia. Indo-Europeans do not 
always win. 




After an extensive criticism of recent theories concerning the origin of the Indo- 
Europeans, Igor Diakonov concluded, 'Perhaps our readers will ask: u but what 
do you think actually happened?" If I knew, I would develop my own theory 
instead of criticizing somebody else's. However, if it is permitted to advance 
some hypotheses in this context, I will offer mine as well.' Diakonov then 
suggested a Southeast European homeland which is incompatible with the 
course of my own arguments. Nevertheless, some attempt at recapitulation is 
called for and it also provides a convenient opportunity to note particular 
problems and possibilities for future research. 

Proto-Indo-European probably evolved out of the languages spoken by 
hunter-fishing communities in the Pontic-Caspian region. It is impossible to 
select which languages and what areas, though a linguistic continuum from the 
Dnieper east to the Volga would be possible. Settlement would have been 
confined primarily to the major river valleys and their tributaries, and this may 
have resulted in considerable linguistic ramification. But the introduction of 
stockbreeding, and the domestication of the horse, permitted the exploitation of 
the open steppe. With the subsequent development of wheeled vehicles in this 
area, highly mobile communities would have interacted regularly with the more 
sedentary river valley and forest-steppe communities. During the period to 
which we notionally assign Proto-Indo-European (4500-2500 bc), most of the 
Pontic-Caspian served as a vast interaction sphere. As Nikolai Merpert has 
observed concerning the social organization of the Yamnaya cultural area, the 
almost instantaneous spread of uniform burial rites over vast areas of the steppe 
suggests movements at the tribal level of society. Words would have passed 
freely between different dialects, and the later isoglosses which seem to leap 
geographical boundaries, such as Greek or German and Tocharian, may have 
been the result of these interactions. In addition, higher versus lower variants of 
Indo-European languages may have been spoken, which would further account 
for why some linguistic groups preserve certain words and others lack such 
reflexes. In the east, both Proto-Indo-Europeans and later ancestors of the 
Indo-Iranians were in contact with Finno-Ugric speakers. In the west, the 
shared agricultural vocabulary of the European languages may have developed 
along the middle Dnieper or in contact with the numerous Tripolyean settle- 
ments of the western Ukraine. 

As the period 4500-2500 BC is an arbitrary segment of the linguistic pro- 
cesses of the Pontic-Caspian, it is difficult to pick and choose which cultures in 
this region might not have been Indo-European. Marija Gimbutas traditionally 
dismisses the Dnieper-Donets culture as genetically different from the Kurgan 
line of cultures since it practised a different economy, buried its dead in group 
pits and displayed a different, more massive, physical type. But since it 
flourished and died during the Proto-Indo-European period, and contributed 
to the formation of the Sredny Stog and possibly Samara cultures, it may well 
have been another component of the overall linguistic continuum which 
evolved into Proto-Indo-European. Similarly, we cannot isolate out languages 



which may have been associated with the Kemi Oba culture of the Crimea, or 
those that might have been connected with north Caucasian cultures, as non- 
Indo-European. The emphasis here is entirely on a territorial sphere of mutual 
interrelationships and that is the best we should ever hope for. 

The eastern border of Proto-Indo-European is insecure, at least in the fifth 
and early fourth millennium bc. It may be that future work east of the Urals will 
uncover cultures genetically related to those of the Volga-Ural region, and that 
we will be impelled to extend our interaction sphere to the east. But, for the 
present, the archaeological picture suggests an expansion eastwards across the 
steppe and forest-steppe of western Siberia beginning in the late fourth 
millennium. Those who reached the far eastern periphery became the 
Afanasievo culture, and both their location and their isolation from other steppe 
tribes indicates that they were well positioned to become the ancestors of the 
Tocharians. What happened to the Afanasievans after the occupation of this 
region by the genetically unrelated Okunevo culture remains a mystery, but as 
they were replaced by a more northerly culture, a retreat southwards towards 
Chinese Turkestan would be a plausible solution. 

After the expansion of Indo-Europeans to the Yenisey, the Indo-Iranian 
languages evolved, in the third millennium bc, in the broad expanse between at 
least the Volga and Kazakhstan. Out of this staging area there was a gradual 
drift southwards through the proto-urban communities of Central Asia. By the 
second millennium, Indo- Aryan was spoken by tribes south of the Caspian, and 
probably also in Afghanistan-north Pakistan from whence it ultimately pressed 
southwards into the Indus Valley. Concurrent with these developments, 
Iranian was evolving on the steppe and was then subsequently carried south 
into present-day Iran and Afghanistan, while the steppe itself was largely left to 
Eastern Iranian-speaking tribes. These expanded as far east as the Yenisey and 
westwards across the Pontic-Caspian where they have left us many of the names - 
of the major rivers north of the Black Sea. 

In bringing the Iranians into the Pontic region in comparatively recent times, 
I indicate that I do not share the opinion of many who have imagined that the 
Indo-Iranians dominated the entire Pontic-Caspian steppe from the Eneolithic 
period onwards. Rather, I prefer to see the Eneolithic cultures speaking largely 
undifferentiated (or at least anonymous) Indo-European, and that the 
Eneolithic expansions into the Balkans were carried out by linguistic ancestors 
of many of the European and Anatolian languages. Hence, the entire Pontic- 
Caspian steppe provided the environment for mutual interactions which would 
explain why we can isolate out Greek-Armenian-(possibly) Thracian-Iranian- 
Indic as an extensive continuum sharing a number of linguistic innovations. To 
place this continuum in the third millennium bc indicates that we must propel 
those Indo-Europeans ancestral to the Anatolian languages out of the Pontic- 
Caspian at an earlier date. A late fifth- or early fourth-millennium expansion is 
what we are looking for, and we have seen that there are two possibilities. There 
is disputable evidence for intrusions from the Pontic steppe into the Caucasus, 
and possibly eastern Anatolia, by the fourth millennium bc. If the evidence 



146 Ait Hons oj the world where an Indo-European language is either primary or recognized as 
an official language of state. 

could be strengthened and traced westwards and southwards into areas where 
we have later record of Anatolian speakers, then it would provide a most 
plausible solution. For the present, however, the evidence for movements 
across Anatolia from the northwest appear somewhat more attractive. This 
would tie the Anatolian dialects to the Early Bronze Age cultures around Troy 
and, by extension, to the Balkan Danubian region. Hence, the earliest evidence 
for steppe intrusions in the Balkans, about 4000 bc, may have been associated 
with Indo-Europeans who carried their language into northwest Anatolia. 
Subsequent linguistic developments involving the ancestors of Greek, Arme- 
nian, Thracian, Phrygian, Illyrian and some of the Indo-European languages 
we later encounter in Italy followed on the continual movement of peoples from 
the steppe into Southeastern Europe throughout the fourth and into the third 
millennium bc. 

When left w ith a choice between an archaeological model that is unconfirmed 
versus one that seems linguistically implausible, I would opt for the former 
hoping, like Miroslav Buchvaldek, that future evidence may rescue it. Hence, I 
w ould have to take it on intuition that some form of historical relationship 
between the Pontic and Central and Northern Europe will eventually be 
demonstrated, even if the evidence today is not conv incing. Either way, it is 
most probable that the Corded Ware horizon provided the vector for a series of 
Indo-European languages that spread both to the west as far as Holland and east 
into the Baltic and upper Volga. Out of this later emerged possibly the Celtic 
and Italic, and more certainly the Germanic, Baltic and Slavic languages. 



The expansion of the Indo-European languages was not completed by the 
Bronze or Iron Ages but has been an ongoing process which continues to the 
present. The rise of Europe and its colonial extensions into Africa, Australia 
and the New World carried English, Spanish, Portuguese and French into new 
lands where they are still assimilating the speakers of native languages. And in 
Asia, Russian is continuing to spread across the eastern territories of the Soviet 
Union. Over a period of 6,000 years we have witnessed the emergence and 
expansion of a language family until it embraces nearly half the population of 
this planet. 




The ancestors of the Aryans cultivated wheat 
when those of the brachycephalics were probably 
still living like monkeys. 

Georges Vacher de Lapolge, 1899 

The Aryan Myth 

We cannot examine the legacy of the Indo-Europeans without first dispelling 
the spectre of the 'Aryan Myth'. The world is all too familiar with how the 
concept of racial supremacy was implemented by the National Socialists in 
Germany, and we would be quite mistaken to imagine that this grotesque 
obsession with the Indo-Europeans or, as they were then more popularly 
known, the Aryans, was merely the creation of a handful of Nazi fanatics. A 
fascination with the 'Aryans' was, in fact, very much part of the intellectual 
environment of the nineteenth and early twentieth centuries. We may dismiss 
Hitler's assertion that only Aryans were the 'founders of culture', but why 
should V. Gordon Childe have been compelled to find them 'fitted with 
exceptional mental endowments' and 'promoters of true progress'? 

The blond, blue-eyed Aryan, fathered in Northern Europe, convinced of his 
superiority and obsessed with his racial purity, was the product of numerous 
intellectual currents of which the development of Indo-European linguistics 
was but one. Leon Poliakov has shown that the roots of this caricature reach 
back into the near-universal longing of the peoples of Europe to secure for 
themselves an illustrious ancestry. The Romans, of course, had sought theirs at 
Troy. By the Middle Ages the aristocracy of Spain were boasting of their 
superior Visigothic blood which set them both apart from and above their 
subjects, while the French endured chronic schizophrenia whether they were 
the linear descendants of Vercingetorix and his Gauls (Celts) or Charlemagne 
and his Franks (Germans). Some English, not content with their obviously 
mixed ancestry of Britons, Anglo-Saxons, Vikings and Normans, felt it 
necessary to drag a lost tribe of Israel to their shores to provide them with a still 
more ancient peerage. The Germans, on the other hand, saw their own history 
begin with those very expansions that provided the Visigoths, Franks and 
Anglo-Saxons - the illustrious ancestors - of their neighbours. When Tacitus 
maintained that the Germans were pure of blood, unmixed with other races and 
autochthonous, there was little reason to deny that their origins lay in Northern 
Europe. If the Church required a Biblical link, then Ashkenaz, a grandson of 
the prolific Japhet, could be found to trek his way to Northern Europe and 

establish the German people. But it was really pride in local origins that 
attracted the Germans, especially during the Reformation, when many found 
themselves pitted against what they despised as the weak, corrupt and foreign 
world of Rome. By the eighteenth and nineteenth centuries, the industrial and 
intellectual explosion of the Germanic-speaking lands further encouraged the 
belief in both German autochthony and the rising destiny of the north. 
Concurrent developments in physical anthropology and comparative linguis- 
tics helped to fuel this belief. 

Coincidental with the rise of anthropology was the development of the 
concept of race. Western scholars, once they had isolated the different races of 
mankind, could hardly resist placing their own Caucasians at the top. Race was 
easily confused with ethnic group, nation and language. History was reduced to 
a succession of races, each holding power for a time before passing the baton 
on to a more energetic race. The busy seaports of England, Holland, Germany 
and Scandinavia, or the intellectual salons of London, Berlin and Vienna left 
little doubt in their own locales where that energetic race dwelt. The superiority 
of the north could even be measured by physical anthropologists who, by the 
nineteenth century, seemed to be ceaselessly measuring the cephalic index. 
Here they distinguished the dolichocephalic (long-headed) Nordics from their 
brachycephalic (broad-headed) neighbours to the south. Surprisingly enough, it 
was not until about 1870 that the blue-eyed blond, previously the caricature 
of the dreamy romantic, became the stereotype of the virile male. A superior 
Nordic physical type had been discovered by science; it remained for the 
philologists to provide him with an ancient and illustrious ethnic identity. 

The discovery of the Indo-European language family did more than simply 
elucidate the historical relationship between many European and Asian 
languages. It severed once and for all the fantasy of deriving all languages from 
Hebrew, and by extension, Adam. The indivisibility of the human race was 
being destroyed not only by those who profited from exploiting different 
peoples, but by science itself. Following the West's discovery of the wealth of 
Indie and Iranian literature, European scholars looked beyond Eden to seek 
their own more illustrious forebears in Central Asia, Iran and India. Although 
Indo-European and Indo-Germanic had both been coined early in the nine- 
teenth century, Max Miiller, and other linguists, encouraged the use of Aryan 
to describe the ancient Indo-Europeans. Naturally, if these early Aryans were 
the ancestors of the Europeans, then they too must have been part of the 
superior white race. 

A peculiar linguistic fallacy arose that provided further proof of Aryan 
superiority. Research into the essential structure of the world's languages 
revealed several basic types depending on how grammatical elements were 
indicated - inflected, agglutinative, and analytic. Although these were 
originally only typological classifications, many could not resist interpreting 
them as various stages in the evolution of language. A 'simple' word-based 
language (analytic) such as Chinese, for example, was placed on the bottom of 
the scale. Languages were then seen to develop to the agglutinative type (as in 


the Altaic and Uralic languages where separate endings are added to nouns) 
and, ultimately, to the inflected type, such as Indo-European. This suggested 
that the Indo-Europeans, or Aryans, having ascended the scale of linguistic 
evolution to its summit, spoke a more 'advanced' language than their 
neighbours. Canon Isaac Taylor, for example, once proposed the notion that 
the Indo-Europeans were essentially 'an improved race of Finns'. 

But even if superiority of physical type, language, and culture were all being 
united under the name of Aryan, there was still one essential element missing - 
with the singular exception of Roger Latham's claim that the Indo-European 
homeland lay in Europe, the common opinion of most scholars prior to the later 
nineteenth century was that the Aryan homeland must lie in Asia. While no one 
doubted that the Aryans belonged to the white race, up until the end of the 
1860s most believed that this race originally dwelt somewhere in the vicinity of 
the Hindu Kush or Himalayas. There was no reason to seek their home in 
Northern Europe. 

It was in the 1860s that a number of scholars, particularly linguists, began to 
publish their doubts about an Asiatic homeland for the Indo-Europeans. Some 
of the reasons adduced, such as the Beech argument, have become part of the 
traditional stock of homeland research. But for our story it was the rise of 
physical anthropology that is most significant, since it eventually came to be 
applied to the problem of Indo-European origins. By 1870 Lazarus Geiger was 
supporting a homeland in Germany, employing the argument that the original 
Indo-Europeans must have been light-skinned blonds. But the man who first 
carried this theory into the widest arena was Theodor Poesche. 

In 1878 Poesche surveyed the historical references to the various Indo- 
European peoples and assembled his evidence to demonstrate that they were 
regularly described as blue-eyed blonds. Even the brahmans of India were 
lighter than the lower castes - an observation made much earlier by Julius 
Klaproth, who had coined the term Indo-Germanic. From this one might 
conclude that the Aryans were originally light-skinned, fair-haired, blue-eyed 
and, in accordance with the anthropological evidence already adduced, 
dolichocephalic. Poesche then sought the centre of all of these characteristics 
and astoundingly enough concluded that the point where one might locate the 
highest incidence of albinism in Europe must have been the centre of the Aryan 
race. At last the Aryans had found a European homeland - in the Pripet marshes 
of Eastern Europe. 

An Aryan homeland in the unhealthy environment of a swamp was hardly 
conducive to the development of the 'powerful, energetic blond race' or so Karl 
Penka argued in 1883. Rather, Penka pressed into service all the disciplines he 
could - archaeology, linguistics, anthropology and mythology - to demonstrate 
that the Aryans originated in southern Scandinavia. One of the arguments 
emphasized by Penka is still very current today. Dfenying any evidence in the 
archaeological record for major intrusions into Northern Europe, Penka 
concluded that the Aryans must have originated there and could not be derived 
from elsewhere. Penka's works, largely a series of polemics, were soon widely 



accepted. Noted anthropologists such as Rudolf Virchow and even Thomas 
Huxley not only concerned themselves with identifying the original Aryan race 
but also concurred that they were originally a race of 'blond dolichocephalics'. 
The great Indologist Max Miiller, annoyed by the madness he had helped to 
create, blasted the anthropologists who spoke of an 'Aryan race, Aryan blood, 
Aryan eyes and hair' as a lunacy comparable to a linguist who spoke of 'a 
dolichocephalic dictionary or a brachycephalic grammar'. But it was too late: 
the Indo-Europeans and racism had become inseparable in the minds of many 
scholars. Although there would always be linguists and anthropologists to 
protest, the superiority of the ancient Aryan Nordic race had entered popular 
political culture. 

The subsequent history of the Indo-European problem until the end of the 
Second World War was played out by several major camps. A European 
homeland was favoured by the great majority of scholars although there were 
always a few who would assert the ancient claims of Asia. Mo$t arguments, 
however, were fought between those who supported a North European 
homeland (the German archaeologist Gustav Kossina and his followers, the 
linguist Herman Hirt, and many others) versus those who supported a steppe • 
homeland, either European or Asian (Otto Schrader, Sigmund Feist, Alfons 
Nehring, Wilhelm Brandenstein, Wilhelm Koppers). Some scholars struggled 
to maintain a middle course, others provided cdhiic relief. The American 
Charles Morris, for example, settled the earliest Ar>^ns in the Caucasus where 
they not only perfected their language, but, in an instance of racial stereotyping 
that tells us far more about the author than the Indo-Europeans, they gained, 
according to Morris, their 'enthusiasm' frym the darjcer races to their south. 
Cokamanya Bal Gangadhar Tilak provided the world with an entire mono- 
graph marshalling all the available mythological evidence to prove that the 
Aryan homeland was the North Pole. 38 This incredible theory gained at least 
one supporter when Georg Biedenkapp, flushed with enthusiasm for Tilak's . # 

hypothesis, produced his own boolcsummarizing the Indian savant's work in 
German and added further evidence of' his own. The Icelandic linguist 
Alexander Johannesson concocted another bizarre theory that related Indo- 
European roots to bird calls (Proto-Indo-European *ker~ was imitative of a 
raven), grurrts, and lo^d natural sounds which, according to him, could best be 
heard on the shores of the Baltic Sea. . 

The myth of Aryan supremacy* somewhat more evident in some pre-war 
anthropological journals than among the linguistic ones, was, in varying 
degrees, a widespread phenomenon until -the consequences of its political 
expression made it anathema in the academic, world. One hardly need 
emphasize that the implementation of Aryan supremacy by the Nazis was # 
wholly inconsistent with Aryan as a linguistic term; Yiddish is as much an Indo- 
European language as any other German dialect, while Romany-speaking 
Gypsies had a far better claim to. the title of Aryans than any North European. 
Thus, the myth of Aryan supremacy was neither a direct nor necessary conse- 
quence of the philological discoveries *of the nineteenth century, but rather 



the misappropriation of a linguistic concept and its subsequent grafting onto an 
already existing framework of prejudices, speculations and political aspirations. 
The Indo-Europeans leave more than the legacy of Aryan supremacy. 

The legacy 

If the development of comparative philology played an unfortunate role in the 
creation of twentieth century racism, it should also be credited with providing 
the tools by which scholars were able to elucidate the cultural relationships and 
origins of the numerous non-Indo-European peoples of the world. The same 
techniques employed to compare the various lexical and grammatical items of 
the Indo-European languages were, and still are, equally applicable to the 
Algonquin, Altaic, Athapascan, Bantu, or indeed any other group of languages. 
Linguists, originally trained in the field of Indo-European, set out to establish 
the relationships between the other languages of the world, to reconstruct their 
proto-languages and investigate their origins. 

When we turn to the actual achievements of the Proto-Indo-Europeans we 
find a rather diff erent legacy from that of most ancient peoples. Unlike Samuel 
Kramer who can list the Sumerian 'firsts', there are few such achievements that 
we can credit to the Proto-Indo-Europeans. Recognizing this, and finding them 
so often in the position of destroyers of earlier cultures, V. Gordon Childe was 
thrown back to extolling their 'excellent language and the mentality it 
generated' while Albert Carnoy relished their expansive nature, their spontane- 
ity, their creativity, their suppleness of spirit and a host of similar virtues. Why 
they should have been given such singular praise is unclear; can one be so 
perverse as to imagine the Chinese as uncreative, the Finns and Hungarians as 
mentally impoverished, or the Turks and Arabs as saddled with second class 

If we must have concrete legacies, then the best claim is that of horse 
domestication and the social consequences this revolution in transportation and 
warfare brought to the world. In addition, the Indo-Europeans are at least one 
of the candidates for the inventors of wheeled vehicles, although a number of 
non-Indo-European peoples have every bit as good a claim. But such instances 
of historical priority hardly constitute the type of legacy that persists uniquely 
among the Indo-Europeans. 

Ideology is often regarded as the central core of culture and it is here that 
some would see the most striking evidence for the Indo-European legacy. We 
have already seen how the Dumezilian school not only acknowledges the 
genetic relationship of the Indo-European languages but also the persistence of 
an inherited ideology. The new comparative mythologists maintain that the 
trifunctional ideology of the Proto-Indo-Europeans permeates the religious 
texts of the ancient Indians and Iranians, emerges in the epic poetry and drama 
of the Greeks, hides behind the facade of history among the early Romans, and 
expresses itself in the prose tales of the medieval Germanic and Celtic peoples. 
Even with that overlay of Judeo-Christian ideology which characterizes 



14; The persistence oj Indo-European tri partition? Peter BrtUgefi sixteenth-century 'Land 0/ 
Cockayne' depicts the clerk, warrior and cultivator, the representatives of the three ' functions of 
Indo-European society. 

European culture, the Indo-European ideological structure still surfaces, as we 
see in the medieval tendency to equate the three sons of Noah - Japhet, Shem 
and Ham - with the three estates of society: nobles (warriors), clerks (priests), 
and serfs (cultivators). C. Scott Littleton has even suggested that this tripartite 
div ision of society extended unconsciously to the framers of the American 
Constitution, who, in dividing the totality of their state into three branches, 
were as much the heirs of their Proto-Indo-European ancestors as Zarathustra 
or the brahmans of ancient India. 

It is natural to query whether social tripartition, and all of its attendant 
expressions in myth, literature, law, medicine, and folklore, is uniquely Indo- 
European. To many, the system proposed by Dumezil seems so natural as to be 
universal, and hardly the specific legacy of an Eneolithic people of Eastern 
Europe. Yet the 'new comparativists' argue that, even if such a system of 
priests, warriors and herder-cultivators seems natural, it is the treatment of this 
structure as a special class of concepts requiring and receiving almost endless 
elaboration in all spheres of cultural ideology and behaviour that makes it truly 
unique to the Indo-Europeans. Littleton has argued that similar methods 
applied to myths of other language groups would reveal other systems of 


ideology ranging from the seven-fold (astral) ideological paradigms of the 
Semites, to the four-fold (directional-based) systems of certain American 
Indian groups. Indeed, one may argue that tripartition itself need not result in 
the Indo-European system. Lawrence Krader, in his study of the non-Indo- 
European Buryat Mongols, observes that their 'triple division of the social 
world' is carried through into the spirit world and the three souls of men. The 
Buryat spirits are arranged according to the three different Buryat social 
classes, aristocrats, commoners and slaves, categories that are tripartite but not 
those of the Indo-Europeans. 

Whether or not one is confident that such an ideological inheritance exists, 
the most secure legacy of the Indo-Europeans is surely to be found in the 
language spoken by over two billion people in the world. It is irrelevant whether 
we regard ourselves as Europeans, Asians, Africans, or Americans, we cannot 
escape this legacy if we speak an Indo-European language. We cannot ask 
questions of where, when, who or how, or answer them with our most basic 
pronouns, we cannot count, refer to the basic parts of our bodies, describe our 
environment, the heavens, basic animals or relatives, or express our most 
fundamental actions, without making frequent recourse to an inherited system 
of speech that our linguistic ancestors shared 6,000 years ago. 

The common linguistic heritage of the Indo-Europeans was only discovered 
in the eighteenth century ^nd it has seldom, if ever, impinged on the behaviour 
of the different Indo-Europeans. History provides little evidence that different 
Indo-European groups ever recognized their mutual kinship. If the ancient 
Greeks disparaged an Indian as a barbaros, the Indian dug into the same 
linguistic legacy to dismiss his non-Aryan neighbours with precisely the same 
word, barbaras. If kinship of the Indo-Europeans was overlooked in the past, it 
hardly needs emphasizing that it is absent today. In characteristic hyperbole, 
Hitler once wrote that the collapse of the Aryans would see the light of 
civilization extinguished in the world; given the distribution of nuclear arms on 
this planet, it is far more likely that it will be Indo-Europeans who will end it 
themselves. Yet we need not finish pessimistically, but rather hope to remind 
the great superpowers, that whatever their political differences, when they 
speak to one another, they do so in words that were once common when they 
shared the same language, the same home and the same beliefs. 

Notes to the text 

1 Jones had a number of predecessors (in addition to Parsons) 
who recognized the affinity of the languages of India and Iran 
with those of Europe. As early as 1583 Thomas Stevens, an 
English Jesuit working in India, wrote that 'many are the 
languages of these places. Their pronunciation is not disagree- 
able and their structure is allied to Greek and Latin'. Two years 
later an Italian merchant Fillipo Sasseti observed that there was 
much in common between Sanskrit.and the European languages 
because in Sanskrit 'we can find many of our nouns, especially 
numbers: the 6, 7, 8, and 9, God, serpent, and others'. By the 
seventeenth century scholars were also accepting the similarities 
between Greek and German, and Franciscus Rapelengius 
argued for the association of German with Persian. The Dutch 
scholar, Marcus Boxhorn grouped Greek, Latin, German and 
Persian under the name of 'Scythian', a theory defended by no 
less a scholar than Leibnitz. Finally, in 1768 the Jesuit 
Coeurdoux derived Sanskrit, Latin, Greek, Slavonic, and the 
other languages of Europe from the language of Japhet (see 
Mukherjee, S. N., Sir William Jones, Cambridge 1968). 

2 Both Rask and Bopp were required to labour against one of 
the last gasps of Goropianism when the great German Romanti- 
cist, Friedrich von Schlegel, published his Oberdie Sprache und 
Weisheit der Indier ('On the Language and Wisdom of the 
Indians', 1808) wherein he proposed that all of the Indo- 
European languages might be derived from Sanskrit itself. The 
fallacy was still very much alive in 1828 when Vans Kennedy 
published his Researches into the Origin and Affinity of the 
Principal Languages of Europe and Asia. Oddly enough, Ken- 
nedy dismissed the idea that Persian and the Celtic languages 
were related to the other Indo-European languages. Besides 
Thomas Young's introduction of the term Indo-European, 
mention should also be made of Julius Klaproth who coined 
'Indo-Germanic', which is most often employed in German- 
speaking lands (indicating the extent of the language family from 
India to the Germanic speakers of the Atlantic). As we will see 
later, Aryan became a popular term for the Indo-Europeans in 
the later nineteenth and early twentieth centuries. In the first 
edition of the Cambridge History of India, P. Giles unsuccess- 
fully attempted to introduce the term 'wiroi' for the Indo- 
Europeans, employing their own (reconstructed) word for 'the 

3 In this text, I employ a traditional system reconstructed for 
Proto-Indo-European that includes the vowels a, e, i, o, u (both 
long and short); the semivowels y and d>; the nasals m and n and 
liquids / and r; the sibilant s; and a rich assortment of consonants 
that are arranged in the following triads: labials, pbbh\ dentals, / 
ddh; velars, k ggh; palatals, 6 g gft\ labio-velars, k w g" g w h. In 
addition to these, there was the indeterminate vowel 3 (schwa) 
which also functioned as a consonant (laryngeal). The existence 
of laryngeals was originally predicted on the basis of the internal 
structure of the Indo-European languages, and this was 
confirmed by the discovery of Hittite which yielded an h-like 
sound in the predicted position. The number of different types 
of laryngeals is still debated and several different systems of 
representation are employed by various linguists. Moreover, the 
past decade has seen major challenges to the traditional 
nineteenth-century reconstructed system (see Szemerenyi, O., 
'Recent developments in Indo-European linguistics', Transac- 
tions of the Philological Society 1985: 1-71, for brief review). 
Purely in the interests of graphic simplicity, the more traditional 
reconstructions are employed in this work. 

4 The personal names include those that are similar to names 
later attested by the Hittites themselves, for example, Sa-li-nu- 

ma- where Hittite salh- means 'great'. Some names such as Ta- 
ak-sa-nu-ma-an appear to be constructed from Indo-European 
roots (Hittite taks- 'make, build', Sanskrit taksan and Greek 
tekton 'carpenter'). 

5 The Hittites adopted the Hattic capital Hattusa and 
personal names, such as Hattusilis, from the Hattic population. 
Among the Hattic loan words in Hittite are words specifically 
associated with the palace polity of the Hatti, for example, 
throne, lord, queen mother, noble, and a whole series of names 
for officials and cult functionaries. Recognizable Hattic loan 
words pertaining to material culture include iron and some 
words for bread (Kammenhuber, A., 'Protohattisch- 
Hethitisches', Miinchener Studien zum Sprachwtssenschaft 14 
(1959): 63-83.). It is generally assumed that the Hattic-Hittite 
bilingual ritual texts were necessary since, although the Hittites 
had apparently adopted the Hattic religion (and others as well), 
Hattic had gone out of use. 

6 Archaic features of Hittite include laryngeals, r/n stem 
nouns, archaic pronomial forms, medio-passives; losses (or 
perhaps absences since a number of linguists would dispute that 
they ever developed in Hittite) include the feminine, the dual, 
the aorist, the optative. A number of the features absent from 
Hittite but which are reconstructed to Proto-Indo-European 
from other languages are also absent from Hattic. Hence a Hattic 
substrate has sometimes been credited as one of the major 
reasons for the restructuring of the Hittite language away from 
that evidenced by the other Indo-European languages. 

7 G. Steiner (1981) has suggested that the Hittites were 
already well-settled in Kanes when the Assyrians arrived, and as 
the dominant population of one of the most important trading 
outposts their language served as a lingua franca across central 
Anatolia, coming to serve as the chancellery language of the 
Hatti. According to Steiner, the Hittites of the Old Kingdom 
were not primarily 'Hittites' but rather of Hattic stock which 
had adopted the new lingua franca. 

8 The horse remains from the Late Eneolithic in eastern 
Anatolia are traditionally identified as wild individuals. S. 
Bokonyi, ('Horses and Sheep in East Europe in the Copper and 
Bronze Ages', in SVSkomal and E. Polome (eds), Proto-Indo- 
European: The Archaeology of a Linguistic Problem, Washington, 
D.C., 1987: 136-144), however, suggests that they were from 
domestic animals, and he associates them with the spread of 
domestic horses from the Pontic-Caspian region. 

9 Linguistically, from what we can tell of Phrygian, it shares 
one apparent archaic feature (medio-passives in r) with Hittite 
and several other 'peripheral' dialects. It also shares the augment 
with Armenian, Greek, and Indo-Iranian, a putatively later 
feature assignable to the 'southern' dialects. This could suggest 
that Phrygian was earlier situated somewhere close to Anatolian 
to avoid later innovations and also close to the Greek- Armenian- 
Indo-Iranian chain of southern languages. Geographically, the 
southern Balkans or northwestern Anatolia should fit. 

10 Regarding Armenian origins, Colin Renfrew (1987: 72) 
dismisses their traditional derivation from Southeast Europe as 
a product of the 'propensity to talk in terms of migrations', and 
argues that 'there is no very clear reason to link the Armenian 
language with western Anatolia or Thrace'. Renfrew fails to 
indicate that / Armenian cannot be derived directly from any of 
the adjacent Anatolian languages (which all lack common Proto- 
Indo-European features found in all the other Indo-European 
languages, including Armenian); and 2 Armenian shares major 
isoglosses that associate it most closely with Greek, and linguists 
generally require that the linguistic ancestors of the Armenians 



and Greeks were either identical or in a close contact relation - 
Thrace or western Anatolia being the most economical explana- 
tions. Obviously, the Armenians could not have arrived in their 
historical seats without migrating since it was previously 
occupied by the non-Indo-European Urartians. 

Renfrew takes some comfort from the fact that Gamkrelidze 
and Ivanov (1985) also place the Indo-European homeland in 
eastern Anatolia on linguistic grounds. While the two Soviet 
linguists do indeed place the homeland on the eastern Anatolia- 
Armenian plateau, they also accept the existence of a 'Greek- 
Armenian community' as an 'independent entity' and are 
consequently forced to send the Greeks migrating from eastern 
Anatolia (the reverse of the traditional model that has the 
Armenians migrate from the Balkans). Diakonov (1985) rightly 
finds this Greek migration across the face of the Anatolian 
languages quite incredible. 

1 1 Renfrew (1987: 1 78-197) rejects as unproven the exclusion 
of an Indo-Aryan identity for the Indus Valley Civilization and 
proposes the hypothesis that the Indo-Aryan languages may 
have extended to west Pakistan with the spread of agriculture 
before 6000 bc. In this hypothesis, the Indo-Aryans, or at least 
their linguistic ancestors, were in the Indus region for thousands 
of years prior to the traditional interpretation of the linguistic 
data. He prudently labels this suggestion 'Hypothesis A' and 
also admits the possibility of the traditional model (Hypothesis 
B) discussed in this book. Other than the reasons cited in the 
main text there are other very serious grounds for rejecting 
Renfrew's Hypothesis A. A model that 'explains' the presence of 
the Indo-Aryans by referring them to the expansion of the 
Neolithic economy from the Zagros region around the seventh 
millennium bc ignores the close linguistic relationship between 
Indo-Iranian and Greek. In Renfrew's model, if the Indo- 
Iranians moved off to the east of what would become Anatolian, 
and the linguistic ancestors of the Greeks moved off to the west, 
there can be no intelligible explanation for how these different 
subgroups share so many isoglosses not found in Anatolian. It is 
curious that for someone who repeatedly adheres to Schmidt's 
'wave model' as a key to explaining the relationships of the Indo- 
European languages, Renfrew pays no attention to the signifi- 
cance of the shared linguistic similarities (isogloss bundles) 
indicated by this model. 

12 Here I follow Pedersen, Crossland, Adrados, Gamkrelidze 
and Ivanov, and others, who regard Tocharian as an archaic 
peripheral dialect. There are, however, many who would 
associate Tocharian much more closely with languages such as 
Germanic or Greek (see Adams, D. Q, 'The Position of 
Tocharian Among the Other Indo-European Languages', 
Journal of the American Oriental Society 104 (1984): 395-402 for 
the most recent review of this problem. Adams himself relates 
Tocharian closest to Germanic and sees no great difficulty 
getting them to their ultimate homes). 

13 Possible textual remains of a pre-Greek language are those 
commonly termed Eteo-Cretan, a small series of inscriptions 
dating from the seventh to the third centuries bc known from 
the island of Crete. They are written in the Greek script but are 
most certainly not Greek nor any other known language and may 
represent one or more of the pre-Indo-European languages of 

14 Although these names have traditionally been interpreted 
as non-Indo-European, a number of linguists have proposed 
Indo-European etymologies, such as Achilleus {*Acht-lawos 
'one who causes distress to the army'). Indo-European etymolo- 
gies for words such as bastleus have been proposed by V. 
Georgiev, one of the main proponents of the Pelasgian theory. 

15 Anna Morpurgo Da vies carried out a cursory examination 
of the Greek vocabulary which revealed less than 40 per cent of 
the lexicon could be ascribed a transparent Indo-European 
etymology, 8 per cent had established non-Greek origins and 
about 52 per cent had no clear etymology (Da vies 1986). 

16 The Linear B script would render all of the following Greek 
words: ago, akos y algos, argos, arkkon, askos y only as ako y and it is 
up to the linguist to determine which, if any, of these words the 
Mycenaean text reflects. Moreover, the syllabic structure of 
Indo-European words, including Greek, tends to be closed, 
while the Linear A and B scripts indicate only open syllables. As 
for the linguistic identity of the earlier Linear A script we can say 
very little. There is sufficient similarity in signs between it and 
Linear B that a number of linguists are confident about 
assigning phonetic values to Linear A inscriptions. They have 
been variously translated as Greek, Anatolian and Semitic; none 
very convincingly. 

17 The entire hypothesis of the Nordwestblock has recently 
been reviewed by W. Meid ('Hans Kuhns "Nordwestblock" - 
Hypothese', Anzeiger d. osterreichischen Akademie d. 
Wissenschaften 121 (1984): 2-21) who, although critical of some 
specifics, indicates that there is sufficient evidence to accept the 
existence of a different language (from Celtic and Germanic) in 
Northwest Europe. Eric Hamp (1987, in Skomal and Polome) 
has been arguing for an apparently non-Indo-European 
substrate language in Northern and Central Europe on the basis 
of the non-Indo-European appearance of an entire series of 
words associated with pigs and pig-breeding that are found 
primarily in the Celtic and possibly the Germanic languages. 
The concept of relics of a non-Indo-European language has 
been explained by some as the linguistic residue of the earlier 
Neolithic settlers of this region. 

18 Names common on Messapic inscriptions have been found 
on tombs in Albania (Toci, Studia Albanica 2 (1969): 163-85). 

19 The model proposed by Renfrew envisages 'an Indo- 
European-speaking population in France and in Britain and in 
Ireland, and probably in much of Iberia also, by before 4000 bc' 
Through the 'wave model' this area gradually emerges as Celtic 
speaking and regionalizes into its various Celtic dialects. This 
scheme for the origin of the Insular Celtic languages is, I believe, 
linguistically most unconvincing. The time depth for these 
Celtic evolutions seems to be totally incongruent with all our 
available linguistic data on the separation of the Celtic lan- 
guages. Our earliest evidence of the Insular Celtic languages 
indicates that / they share common 'Late Bronze Age-Iron Age 
vocabulary' with each other and Continental Celtic, such as 
words for iron, lead, weapons, and chariot terminology, which 
seems to indicate little major dialectal separation from the time 
of their putative settlement (according to Renfrew) to the Late 
Bronze Age or Iron Age, that is about 3,000 years or more; 2 that 
the Celtic languages were extremely similar to one another when 
they first appear in written sources - a phenomenon best 
explained by a relatively recent divergence across Common 
Celtic; j that they began to undergo extreme restructuring 
during the first millennium a d - a process which may be due to 
internal forces but is certainly in conformity with one's 
expectation of a series of languages recently spread over 
linguistic groups speaking radically different languages on the 
periphery of Europe. For what it is worth, traditional Irish 
accounts of their own origins regularly placed them at the end of 
a series of migrations to Ireland, and they identified certain 
elements of Irish population as native in contrast to the Goidelic 
(Irish) speakers (cf. Greene 1983, Piggott 1983, Mallory 1984, 
and MacEoin 1986.) 



20 Renfrew's ( 1 987, 77-86, 1 03- 1 04, 1 09-1 10), discussion of 
the use of linguistic palaeontology as a tool for reconstructing the 
culture of the Proto-Indo-Europeans would appear to discredit 
the entire method. His objections include the following: 

/ A new invention such as the 'wheel' will often carry its 
original name as it is spread to other populations, so if we find 
this word among several languages it does not necessarily tell us 
much about their origins. The observation that a new invention 
might well diffuse with its name attached to it is quite true, and 
historical linguists frequently take advantage of such phenom- 
ena to elucidate the historical relations between peoples. For 
example, linguists know that in their early contacts with one 
another, the Latins borrowed a series of words relating to 
chariotry from their Celtic neighbours in northern Italy. Hence, 
Celtic *karros (Old Irish, Middle Welsh, Old Breton carr) was 
lent into Latin as carrus, a word that is clearly distinguished from 
the native Latin chariot word currus which is derived ultimately 
from the same Indo-European root. The word was then 
subsequently borrowed through Norman French {carte) into 
Middle English as carre, modern English car. The distinction 
between which words are inherited from an earlier linguistic 
stage, and which were borrowed later, is clearly traceable. When 
historical linguists reconstruct the wheel to Proto-Indo-Euro- 
pean, they do so because its outcome in the various Indo- 
European languages conforms to the same rules as all other 
words reconstructed to Proto-Indo-European. That it may have 
been invented by the Sumerians and then borrowed by the 
Indo-Europeans is interesting but irrelevant since the word 
would still be reconstructible to that stage in the linguistic 
continuum that we would designate Proto-Indo-European. 
What is important is that we have no reason to believe that this or 
any of the other vehicle-related terms reconstructed to Proto- 
Indo-European were loan words passing from one differentiated 
group to another (see also note 25). I can hardly claim that there 
do not exist considerable difficulties in assessing the inherited 
versus the borrowed status of some words, especially when their 
phonetic shape does not result in markedly different outcomes in 
different Indo-European languages, but there are procedures by 
which one can make an intelligent evaluation of the data. 

2 The meanings of words change such that we cannot be 
sure of the original meaning of words that we reconstruct; Any 
serious linguist engaged in cultural reconstruction does attempt 
to assemble as much data as possible to establish the most likely 
original meaning. Renfrew cites several rather muddled exam- 
ples of how historical linguists may err in their reconstructions. 
He cites J. Fraser's observation that because we have words 
corresponding to our 'mead', and verbs in Greek and Sanskrit 
meaning 'to be intoxicated', we cannot assume that the Indo- 
Europeans had an intoxicating drink. In actual fact, the root 
*medhu is attested in most 'branches' of Indo-European. In 
Celtic and Germanic it only means 'mead', in Indie and Baltic it 
means 'honey' or 'mead', in Avestan it means 'an alcoholic 
drink', in Greek it means 'wine', in Tocharian and Slavic it 
means 'honey'. We have verbal forms from the same root in 
several branches that indicate the meaning 'to be drunk'. We 
also have another word for honey that does not involve the 
meaning mead. What's the problem? Renfrew cites Fraser's non 
sequitur that because Slavic pivo 'beer' is cognate with Latin 
bibere 'to drink' this illustrates how commonly one may transfer 
the meaning 'drink' to an 'alcholic drink'. If one were to attempt 
to reconstruct a Proto-Indo-European 'beer' on such evidence, 
there would be very good reason to object, however, this has 
absolutely nothing to do with the line of evidence that supports 
the existence of a Proto-Indo-European 'mead'. Other exam- 

ples, similarly inappropriate or out of step with present thinking, 
are trotted out, the solidly reconstructed mixed with the 
debatable. He maintains that 'when we find words related to 
"birch" and "beech" in several Indo-European languages it 
does not follow that the common word in Proto-Indo-European 
from which they were descended had the same specific 
meaning'. The first word, birch, is attested as Proto-Indo- 
European *bhergo and has cognates in Slavic, Baltic, Italic, 
Germanic, Indie and Iranian. It means 'birch' in all these 
languages except Latin where it has been shifted to 'ash', 
plausibly enough because the birch is generally absent from the 
Mediterranean area. Furthermore, the root is paralleled by a 
verbal root meaning 'to shine, become white', again totally 
congruent with reconstructing the meaning as 'birch'. The 
original meaning of the beech word on the other hand, as the 
reader will see from the main text, is very much disputed and 
cannot be regarded as convincingly reconstructed to Proto- 
Indo-European. The lesson here, I believe, is not that the 
comparative method cannot be utilized to reconstruct past 
cultural items, that is, that the birch should be tossed onto the 
same bonfire as the ambiguous 'beech' word, but rather that 
linguists are required to exercise the same sort of judgment 
about their data as one would expect of an archaeologist. No 
linguist could pretend that all lexico-cultural reconstructions 
have been carried out with necessary rigour or that extremely 
naive use has not been made of these reconstructions; however, 
historical linguists are not really so hopelessly lost as Renfrew's 
discussion seems to portray them (see Richard Diebold, 
'Linguistic Ways to Prehistory', in Skomal and Polome, 1987: 

21 The Albanian evidence throws suspicion on reconstructing 
*bhagos as 'beech'. Although it is the most common deciduous, 
tree in the Albanian forests, the Albanians employ the ash word 
{ah from *ok$o-) to name the beech. U*bhagos originally meant 
'beech' it is a bit difficult to understand why the Albanians 
applied the 'beech- word' to the chestnut-oak and then substi- 
tuted the ash-word for the beech (Huld, M. E., KZ 95 (198 1): 

22 In a recent re-examination of *uksan 'ox', Stefan Zimmer 
(KZ 1981, 84-91) emphasizes that this word, attested in Indo- 
Iranian, Tocharian, Germanic and Celtic, regularly indicates 
the castrated ox, which is further indication of its use as a 
draught animal. 

23 Eric Hamp (IF 85 (1980), 35-42) has suggested that the 
word for dog was *pkuon (*pek-kuon 'sheep-dog'?). 

24 Szemerenyi (1977, 96-99) derives Proto-Indo-European 
*wik which yields Indie vis- 'settlement, house, clan, tribe' etc. 
from a Proto-Indo-European verb *weik 'go, march', thus 
indicating that the word originally referred to a group on the 
move (cf. English gang from an obsolete verb gang 'to go') and 
was only later applied to the actual settlement of the clan. 

25 Renfrew (1987, 86) dismisses the use of linguistic 
palaeontology to reconstruct such cultural items as wheeled 
vehicles for the Proto-Indo-Europeans. He writes: 'Certainly, 
the circumstance that the Sanskrit word for "chariot" rat ha, is 
agreed by competent linguists to be cognate with the Latin for 
"wheel", rota y is interesting, and merits historical explanation. 
But that is a far cry from saying that the two cognate words tell us 
that some hypothetical Proto-Indo-Europeans used chariots 
with wheels (or indeed carts with wheels) in their original 
homeland.' The exclusion of such late cultural terms is 
necessary since Renfrew's hypothesis requires an expansion of 
the Indo-Europeans c. 3,000 years before our earliest evidence 
for wheeled vehicles. For the record, the *rota- word cited above 



is attested in I ndo- Iranian, Italic, Celtic, Baltic and Germanic. 
Another cognate word for wheel/wagon is known in Celtic, 
Germanic, Baltic, Slavic, Tocharian, Indo-Iranian, Greek and 
Phrygian. Yet a third is attested in Hittite and Tocharian. A 
word for the shaft or pole of a wagon is known in Hittite, Indie, 
Greek and Slavic. Harness is known in Hittite and Sanskrit. 
Axle is known in Indie, Greek, Latin, Celtic, Baltic, Slavic, 
Germanic and Tocharian. A fairly ubiquitous term for navel in 
the Indo-European languages is also frequently applied to the 
nave of a wheel. The word for yoke, normally associated with 
paired draught, is attested in Hittite, Indie, Greek, Latin, 
Germanic, Baltic, Slavic, Celtic, and Armenian. A verb indicat- 
ing to ride or travel by vehicle is attested in Indie, Iranian, Greek, 
Italic, Baltic and Germanic. The reconstruction of wheeled 
vehicles to Proto-Indo-European is universally recognized by 
linguists and is not based, as Renfrew seems to imply, solely on a 
Sanskrit-Latin correspondence. 

26 Friedrich (1966), Wordig (1970), and Gates (1971) all 
argue that Proto-Indo-European kinship was of the Omaha III 
type; Huld (1981) suggests that it was Omaha II or IV. Beekes 
(1976) maintains that there is no solid evidence for an Omaha- 
type kinship system in Proto-Indo-European. Similarly, 
Szemerenyi (1977) rejects the Omaha classification and, like 
Beekes, emphasizes the close role of ego with mother's brother 
(avunculate) without accepting the terminological identities, for 
example, FaFa-MoBr, postulated for the Omaha system. It is 
possible to assume confidently that the Proto-Indo-Europeans 
evidence a patrilineal system and that in the history of some 
subgroups there is evidence for a confusion of terms across 
generations, for example, FaFaand MoBr or S0S0 with SiSo. Of 
the classic textbook kinship systems, the Indo-European 
languages suggest a system probably closer to Omaha than any 
other; however, it falls far short of replicating the classic Omaha 
system with its series of skewing, merging and half-sibling rules. 
Debate revolves particularly about the issue of whether this 
terminological mixing of generations in some Indo-European 
subgroups reveals traces of an original system which represents 
at least one set of characteristics (of a constellation of other 
features nowhere clearly argued) for an Omaha kinship type, or 
whether it represents later developments within the various 
Indo-European subgroups as their own kinship systems evolved 
to accommodate changing social relationships. It is useful to 
remember that other than an association with patrilineality (a 
feature which we could ascribe to Indo-European kinship 
without this debate), identifying one's kinship as Omaha does 
not appear to provide us with any further information about the 

27 Szemerenyi (1977, 1 25-1 49) provides a thorough summary 
of all the arguments concerning the word arya- and concludes 
that it is not even Indo-European but a Near Eastern, probably 
Ugaritic, loan word meaning 'kinsman, companion'. 

28 Another approach to the dating of Proto-I ndo-European is 
glottochronology (or lexicostatistics) which may be likened to 
the linguistic version of radiocarbon dating. It is based on the 
assumption that languages that once shared the same genetic 
ancestor will diverge from one another through the replacement 
of a basic vocabulary at a constant rate. A measurement of the 
shared or replaced words from a list of the basic vocabulary 
(usually a list 'of 100 or 200 words) is then computed into 
approximate calendar years. The technique, at least for deter- 
mining the time depth of linguistic separations, has not been 
generally accepted by linguists and it has been challenged both 
on the basis of its theoretical premise (that languages experience 
a constant rate of vocabulary 'decay') and the very real 

difficulties of its practical execution. When applied to the 
separation of the Indo-European languages from one another 
(see Swadesh, M, 'Unas Correlaciones de Arqueologia y 
Linguistica', in P. Bosch Gimpera, El Problema Indoeuropeo, 
Mexico, i960; also H. Wittman, IF 74 (1969): 1-10) the results 
normally suggest that the earliest divergences within the Indo- 
European family began about 4500-3 500BC. Many, if not most, 
linguists prefer to exercise their own intuitive sense of time 
reckoning for Proto-Indo-European based on the observed 
differences through time of historically attested dialects such as 
the Romance languages. 

29 The r/n stems are heteroclitics, that is, the stem of the noun 
alters from an r in the nominative-accusative cases to an n in the 
other cases, for example, Hittite nominative wader 'water' but 
genitive wedenas. Traces of these are found in other Indo- 
European languages, such as Sanskrit yakrt 'liver' but genitive 
yaknas, or Latin tecur but iecinoris. Normally, the other Indo- 
European languages levelled the paradigms off according to a 
single form, for example, Lithuanian vasara 'summer' but Old 
Church Slavonic vesna 'spring', from a Proto-Indo-European 
*wes-rfn-. Only in Hittite were the r/n stem nouns still 
productive and not remnant archaisms. 

30 I have intentionally omitted from the 'homeland' discus- 
sion all those hydronymic systems that embrace large portions of 
Europe as too suspect to warrant extended consideration in the 
main body of this text. These begin with Jan Rozwadowski's 
attempt in 191 3 (Rocznik Slamistyczny 6: 39-73) to define an 
'Old European' seat for the Indo-Europeans in Northern and 
Eastern Europe on the basis of the Indo-European etymologies 
for many of the major rivers in these regions. Perhaps the most 
famous 'system' was that of Hans Krahe's 'Alteuropaisch' (Old 
European) river names (see Saeculum 8 (1957): 1-16; also Unsere 
Altesten Flussnamen, Wiesbaden) that spanned Europe from the 
Atlantic to the Baltic and were allegedly formed prior to the 
emergence of Celtic, Germanic, Italic, Venetic and Messapic at 
a time depth of approximately 1 500 bc. W. P. Schmid (see IF 77 
(1972): 1 -1 8, and most recently in Skomal and Polome (1987)) 
has argued that these river names were established even earlier, 
prior to the differentiation of all the Indo-European languages, 
and that they could best be localized to the Baltic region. One of 
his 'proofs' for the Proto-Indo-European age of his names is that 
some can only be provided a meaning by appealing to words in 
individual Indo-European languages outside their area of 
occurrence, for example, Baltic river names such as Indus, 
Indura, Indra are only explainable by reference to Sanskrit indu- 
' drops', therefore, Schmid argues that the river was named in 
the Baltic region before the divergence of the Indo-European 
languages. I find such conclusions remarkable since the Sanskrit 
word indu- 'drops' is described as 'without certain etymology' in 
the standard etymological dictionary of Sanskrit, and even 
where it is provided with an Indo-European etymology, such as 
Proto-Indo-European *oid- 'swell', this would surely suggest an 
independent development in Indo-Aryan and there can be no 
claim to Proto-Indo-European status for such a river name. On 
the basis of such evidence I would have thought that the Baltic 
river names lacked any convincing etymology, and do not 
preserve a Proto-Indo-European word. The roots of many of 
Krahe's river names are often indistinctive («r-, wr-, nar- y 
50/-, and so on) and have also been employed to show the 
presence of non-Indo-Europeans in Iberia and even Dravidian 
speakers in Europe! Hans Kuhn (see KZ 71 (1954): 129^161; 
also Anzetger fur Deutschen Altertum und Deutsche Liter atur 78 
(1967): 1-22) who presented a detailed criticism of all this 
evidence, has himself introduced a second 'Old European' 



system of river names (see Namn och Bygd 59 (197 1): 1-22). His 
alternative system is built of different elements from the first 
which supposedly represented the residue of non- Indo-Euro- 
pean substrate names, especially in Northwest Europe, but 
extending all the way from Anatolia to Ireland. See D. P. Block, 
Namn och Bygd 59 (1971 ): 1 49-1 61 , for a critical review of some 
of these ideas. 

31 The place of the accent in the Indo-European languages 
may be fixed on a certain syllable, usually the first syllable in 
Germanic, Czech and Irish; the penultimate syllable in Polish 
and Latin; or it may be 'free' and move from one syllable to 
another depending on a given grammatical form, for example, 
Sanskrit, Lithuanian, some Slavic languages and Greek. The 
free accent is seen in the nominative, accusative and genitive 
forms for the word for 'foot' in Sanskrit and Greek: 

Sanskrit Greek 
nominative pat pous 

accusative padam poda 

genitive padas podos 

32 'Linguistic palaeontology' was first coined by Adolphe 
Pictet in 1859. His horrendously uncritical treatment of the 
lexical material (he assumed a priori that the homeland was in 
Bactria), plus many subsequent uncritical uses of the linguistic 
evidence, brought the name into disrepute. Other terms for 
describing the technique of reconstructing the proto-culture of a 
linguistic group through the comparative method include 
'Worter und Sachcn' (words and things), iexico-cultural 
reconstruction' or, most recently, 'interpretive etymology'. See 
Diebold (1987) for a review of linguistic ways to prehistory. One 
should note that this technique has been widely applied outside 
Indo-European, for example, with Uralic, Semitic, Athapaskan 
and Algonquin. 

33 See, for example, J. Jorgensen, Saltsh Language and 
Culture: A Statistical Analysis of the Internal Relationships, 
History and Evolution, The Hague, 1969. On a broader scale, a 
significant degree of correlation between major Amerindian 
groups and North American culture areas can be observed (see 
H. Driver, Indians of North America, 1969) although there are 
also major non-correlations, especially evident among linguistic 
groups that have spread over considerable distances, such as 
Athapascan, Uto-Aztecan. 

34 Renfrew (1987) presents two models of I ndo- Iranian 
origins. Model A derives them directly from the Neolithic 
populations of Western Asia while Model B takes them from the 
steppe (see note 11). 

35 The economic foundations of the Seraglazovo culture 
remains problematic. There is some evidence for domestic 
animals in the Caspian depression and on the west bank of the 
Volga, however, an intensive survey on the east in the Volga- 
Ural interfluve has recently uncovered well-preserved settle- 
ments that yielded only wild faunas (Igor Vasiliev, personal 

36 The relationship between the Caspian and the Volga-Ural 
region is a perennial topic of debate among Soviet archaeolo- 
gists. Witness most recently how G. Matyushin (in Zvelebil 
(1986): 133-150) links the southern Urals with the south 
Caspian by way of / their geometric lithic industry (Yangelsk 
culture of Urals; Belt, Hotu, Shanidar, Karim Shahir, Jarmo); 2 
ovicaprids (most southern Ural Neolithic sites; north Mesopo- 
tamia, northern Iran); and j Mediterranean physical type 
(Mullino II in southern Urals; basic population known from 

south Caspian). According to Matyushin all of these contacts 
long preceded the seventh millennium BC. E. Kuzmina (1986), 
however, rejects the derivation of the steppe economies from the 
south Caspian pointing out the unfavourable climatic conditions 
prevailing between the two zones, the absence of domestic 
livestock in the east Caspian Kelteminar culture which would 
have been contemporary with these supposed 'southern' im- 
pulses, and the meagre remains of ovicaprids in the frequently 
cited south Caspian sites such as Djebel Cave. Kuzmina assumes 
that the impulses producing the Neolithic economy were 
ultimately derived from the Balkan-Danubian region. I. 
Vasiliev (personal communication) has recently indicated that, 
although connections between the south Caspian and steppe 
region are not to be denied, the primary orientation of the 
populations north of the Caspian have been with the Caucasus 
and Pontic regions from the Mesolithic through to the Bronze 

37 J. Makkay (in Skomal and Polome (1987)) has recently 
offered a new defence of the LBK theory. He argues that it 
displays the necessary Neolithic economy, continuity over space 
and time suggesting a common language, correlation with the 
area of Old European hydronyms and avoidance of areas of 
suspected non-Indo-European substrates (Mediterranean, At- 
lantic Europe), is strategically situated to explain Indo-Euro- 
pean expansions across the Alps into Italy, offers a donor region 
for the spread of Indo-European languages into Northern 
Europe (TRB), shows no evidence of having been invaded from 
elsewhere, and so on. He argues that the European and Asian 
(I ndo- Iranian) languages had been separated before the Neo- 
lithic or at its very beginning because of the absence of 
agricultural terms in the Indo-Iranian languages. He draws the 
dividing line along the Dnieper and accepts those theories that 
derive domesticated animals in the steppe zone either from the 
Caucasus or the Caspian. The spread of Tocharian (with its 
centum relations in Europe) to Asia is regarded as a mystery. 
Specific objections to this theory include: / Indo-Iranian 
languages do reflect all the domestic livestock terms, names for 
secondary products and even words like 'plough' which are 
clearly associated with agriculture. 2 If the domestic animals 
entered the steppe from the Caucasus or the Caspian, why does 
Indo-Iranian share precisely the same terms with those attrib- 
uted to the Balkan-Danubian zone? Makkay realizes rightly that 
we cannot project the Balkan-Danubian cultures east of the 
Dnieper to bring them into an historical relationship with the 
steppe peoples but he fails to generate an acceptable alternative 
that would explain the Indo-Iranians. A solution to the Indo- 
European homeland problem that cannot explain the Indo- 
Iranians is not a really viable solution. 

38 Tilak's 'polar theory' for Aryan origins was not a bizarre 
quirk of a single individual but rather the culmination of an 
extremely long tradition of analysis of Indo-Aryan myth, for 
example, poems that indicate a home in the north where a day 
and a night lasted six months each, the Pole star rises to the 
zenith, and so on. A modern review of this 'northern cycle' of 
myths can be found in Bongard-Levin (1980) who argues that 
Indo-Aryan, Iranian and Scythian traditions (and by cultural 
contact also Greeks) all shared a common mythology of a 
northern mountainous land which, he argues, could only have 
been acquired in their prior common home on the Pontic- 
Caspian steppe. 



A general reader wishing to acquire a more fundamental basis in 
the structure of the Indo-European languages might wish to 
consult Baldi, P., An Introduction to the Indo-European Lan- 
guages y Carbondale, 111., 1983, or Lock wood, W. B., Indo- 
European Philology, London, 1969. The standard general survey 
of the various Indo-European languages is Lockwood, W. B., A 
Panorama of the Indo-European Languages, London, 1972. 
There are two basic comparative Indo-European dictionaries: 
Buck, C. D., A Comparative Dictionary of the Indo-European 
Languages, Chicago, 1949, arranges the data by subject matter, 
while the basic tool of linguists is Pokorny, J., Indo-Germanisches 
Etymologise hes Worterbuch, Bern, 1959, which is arranged 
according to Proto-Indo-European root. There are a number of 
handbooks on Indo-European culture but the most extensive 
and still quite useful is Schrader, O. and A. Nehring, Reallexikon 
der Indogermanischen Altertumskunde, 2 vols, Berlin and Leip- 
zig, 191 7-1929. A good introduction to the ideas of Georges 
Dumezil and other comparative mythologists is Littleton, C. S., 
The New Comparative Mythology, Berkeley and Los Angeles, 
1982. For a brief review of many of the theories concerning the 
Indo-European homeland see Mallory, J. P., 'A short history of 
the Indo-European problem', JIES 1 (1973): 21-65, while 
Scherer, A., Die (Jrheimat der Indo-Germanen, Darmstadt, 
1968, offers a collection of previous works on the problem. An 
annual bibliography of all works related to the Indo-European 
languages (on the order of 2,000 entries per year!) is provided in 
the journal Die Sprache. 

The references below were selected either on the basis of their 
fundamental importance in Indo-European studies or because 
they provided primary information utilized in preparing this 
book. In the interest of brevity, and excepting instances where 
authors are specifically mentioned in text, most basic conference 
proceedings and multi-authored collections, especially Russian 
and Ukrainian, are listed only by editor and title. 
Abbreviations CAH - Cambridge Ancient History, IF - 
Indogermanische Forschungen,JIES -Journal of Indo-European 
Studies, KZ - Zeitschrift fur Vergleichende Sprachwissenschaft, 
SA - Sovetskaya Arkheologiya 

Anttila, R., An Introduction to Historical and Comparative 

Linguistics, New York, 1972. 
Baldi, P., An Introduction to the Indo-European Languages, 

Carbondale, 111., 1983. 
Delbriick, B., Introduction to the Study of Language, 

Amsterdam, 1882. 
Lockwood, W. B., Indo-European Philology, London, 1969. 
, A Panorama of Indo-European Languages, London, 


Meillet, A., Introduction a t Etude Comparative des Langues 

Indo-Europeennes, Paris, 1922. 
Pedersen, H., The Discovery of Language, Bloomington, 

Indiana, 1931. 

Robins, R. H., A Short History of Linguistics, London, 1979. 
Szemerenyi, O., Einfuhrung in die Vergleichende 

Sprachwissenschaft, Darmstadt, 1970. 
, 'Recent Development in Indo-European Linguistics', 

in Transactions of the Philological Society (1985): 1— 71. 

Adrados, F., 'The Archaic Structure of Hittite: The Crux of 

the Problem', JIES 10 (1982): 1-35. 
Carruba, O., 'Origini e Preistoria degli Indo-Europei 

d' Anatolia', Revtsta di Filologta 97 (1969): 1-30. 

Crossland, R. A., 'Immigrants from the North', CAH 1/2 

(1971): 824-876. 
Friedrich, J. et al., Altklemasiatische Sprachen, Leiden, 1969. 
MacQueen, J., The Hittites and Their Contemporaries in Asia 

Minor, London, 1986. 
Mellaart, J., 'Anatolia and the Indo-Europeans', JIES 9 

(1981): 135-149- 
Puhvel, J., 'Dialectal Aspects of the Anatolian Branch of 

Indo-European', in H. Birnbaum and J. Puhvel (eds), 

Ancient Indo-European Dialects, Berkeley, 1966: 235-247. 
Singer, I., 'Hittites and Haitians in Anatolia at the 

Beginning of the Second Millennium bc', JIES 9 (1981): 

1 19-134. 

Steiner, G., 'The Role of the Hittites in Ancient Anatolia', 

JIES 9 (1981): 150-173. 
Winn, M., 'Thoughts on the Question of Indo-European 

Movements in Anatolia and Iran', JIES 2 (1974): 

1 17-142. 

, 'Burial Evidence and the Kurgan Culture in Eastern 

Anatolia c. 3000 bc: An Interpretation', JIES 9 (1981): 

Yakar, J., 'The Indo-Europeans and Their Impact on 

Anatolian Cultural Development', JIES 9 (1981): 94-112. 

Barnett, R. D., 'Phrygia and the People of Anatolia in the 

Iron Age', CAH 2/2 (1975): 417-442. 
Bittel, K., Grundziige der Vor- und Fruhgeschichte 

Kleinasiens, Tubingen, 1963. 
Haas, O., Die Phrygischen Sprachdenkmaler, Sofia, 1966. 

Diakonov, I. M., Predystoriya Armyanskogo Naroda, Erevan, 

, 'Hurro-Urartian Borrowings in Old Armenian', 

Journal of the American Oriental Society 105 (1985): 

Greppin, J., 'Hittite-z(a), Armenian z-, and the Theory of 

Armeno-Hittite Loan Words', JIES 3 (1975): 87-94. 
Indo- Iranian (general) 

Asimov, M. S. et al. (eds), Ethnic Problems of the History of 
Central Asia in the Early Period, Moscow, 1981. 

Bongard-Levin, G. M., The Origin of Aryans, New Delhi, 

Grantovsky, E. A., '"Seraya keramika", "raspisnaya 

keramika" i indoirantsy', in Asimov et al. (eds), Ethnic 

Problems of the History of Central Asia in the Early Period, 

Moscow, 1 981: 245-273. 
Gupta, S. P., Archaeology of Soviet Central Asia and the 

Indian Borderlands, Delhi, 1979. 
Jettmar, K., 'Die Steppenkulturen und die Indoiranier des 

Plateaus', Irama Antiqua 9 (1972): 65-93. 
Kuzmina, E. E., 'O Nekotorykh Arkheologicheskikh 

Aspektakh Problemy Proiskhozhdeniya Indoirantsev', 

Peredneaziatsky Sbornik 4 (1986): 169-232. 
Mandelshtam, A. M., Pamyatniki Epokhi Bronzy v Yuzhnom 

Tadzhikstane, Leningrad, 1968. 
Masson, V. M. and V. Sarianidi, Central Asia, London and 

New York, 1972. 

Ghirshman, R., Ulran et la Migration des Indo- Ary ens et des 

It aniens, Leiden, 1977. 
Littauer, M. A. and J. H. Crouwel, Wheeled Vehicles and 

Ridden Animals in the Ancient Near East, Leiden, 1979. 
Mayrhofer, M., Die Indo-Arter im Alten Vorderasien, 



Wiesbaden, 1966. 

, Die Arier im Vorderen Orient-ein Mythos?, Vienna, 


Piggott, S., The Earliest Wheeled Transport, London, 1983. 
Indo- Aryan 

Allchin, B. and R., The Rise of Civilization in India and 

Pakistan, Cambridge, 1982. 
Burrow, T., The Sanskrit Language, London, 1955. 
, 'The Proto-Indoaryans', Journal of the Royal Asiatic 

Society (1973)* 123-140. 
McAlpin, D., 'Proto-Elamo-Dravidian: The Evidence and its 

Implications', Transactions of the American Philosophical 

Society 71 (1981): pt. 3. 

Diakonov, I. M., 'Media', Cambridge History of Iran 2 

(1985): 36-148. 
Frye, R., The History of Iran, Munich, 1984. 
Ghirshman, R., L'Iran et la Migration des Indo-Aryens et des 

It aniens, Leiden, 1977. 
Winn, M. M., 'Thoughts on the Question of Indo-European 

Movements in Anatolia and Iran',7/£S 2 (1974): 

1 17-142. 

Young, T. C, 'The Iranian Migration into the Zagros', Iran 


Heine-Geldern, R., 'Das Tocharerproblem und die 

Pontische Wanderung', Saeculum 2 (1951): 225-255. 
Lane, G., 'On the Interrelationship of the Tocharian 

dialects', in H. Birnbaum and J. Puhvel (eds), Ancient 

Indo-European Dialects, 1966: 213-233. 
, 'Tocharian: Indo-European and non-Indo-European 

relationships', in G. Car dona (ed.) et al., Indo-European 

and Indo- Europeans, Philadelphia, 1970: 73-88. 
Liu-Mau-Tsai, Kutscha und seine Beziehungen zu China vom 

2 Jh. his zum 6 Jh. n. Ch., Wiesbaden, 1969. 
Pulley bank, E., 'Chinese and Indo-Europeans', Journal of the 

Royal Asiatic Society (1966): 9- 39. 
— — , 'The Chinese and their neighbors in prehistoric and 

early historic times', in D. Keightly (ed.), The Origins of 

Chinese Civilizations, Berkeley and London, 1983: 

41 1-466. 
Europeans (general) 

Geipel, J., The Europeans, London, 1969. 

Arditis, Elly (ed.), Acta of the 2nd International Colloquium 

of Aegean Prehistory: The first arrival of Indo-Europeans in 

Greece, Athens, 1972. 
Best, J. G. P. and Y. Yadin, The arrival of the Greeks, 

Amsterdam, 1973. 
Cadogan, G., The End of the Early Bronze Age in the 

Aegean, Leiden, 1986. 
Chad wick, J., 'The prehistory of the Greek language', CAH 

2/2 (1975): 805-819. 
Crossland, R. A. and A. Birchall (eds), Bronze Age 

Migrations in the Aegean, London, 1973. 
Davies, A. M., 'The Linguistic Evidence', in G. Cadogan 

(ed.), The End of the Early Bronze Age in the Aegean, 

Leiden, 1986: 93-123. 
Haley, J. and C. Blegen, 'The Coming of the Greeks', 

American Journal of Archaeology 32 (1928): 141-154. 
Hammond, N. G. L., Migrations and Invasions in Greece and 

Adjacent Areas, Park Ridge, N. J., 1976. 
Hausler A., 'Die Indoeuropaisierung Griechenlands nach 

Aussage der Grab- und Bestattungssitten', Slovenska 

Archeologia 24-1 (1981): 59-66. 
Hester, D. A., 'Pre-Greek Place Names in Greece and Asia 

Minor', Revue Hittite et Asianique 15 (1957): 1 07-1 19. 
, 'Pelasgian - A New Indo-European Language?', 

Lingua 13 (1964): 335-384- 
, 'Recent Developments in Mediterranean "Substrate" 

Studies', Minos 9 (1968): 219-235. 
Hiller, S., 'Zur Frage der Griechischen Ein wanderung', 

Mitteil. d. osterreich. Arbeit sgemeinsc haft f Ur- und 

Fruhgeschichte 32 (1982): 41-48. 
, 'Zur Ethnogenese der Griechen', Symposium: 

Ethnogenese Europdischer Vblker, Mainz, 1982. 
Hooker, J. T., 'The Coming of the Greeks', Historia 15 

(1976): 129-145. 

, Mycenaean Greece, London, 1976. 

Merlingen, W., 'Fair Play for "Pelasgian"', Lingua 18 

(1967): 144-167. 
Mylonas, G., 'The Luvian Invasions of Greece', Hesperta 31 

(1962): 284-309. 
Palmer, L. R., The Greek Language, London, 1980. 
Rutter, J. B., Ceramic change in the Aegean Early Bronze Age 

(Occasional Paper No. 5, Institute of Archaeology, 

UCLA), Los Angeles, 1979. 
, 'Fine Gray-Burnished Pottery of the Early Helladic 

III Period: The Ancestry of Gray Minyan', Hesperia 52 

(1983): 327-355- 
Sakellariou, M. B., 'Linguistic and Ethnic Groups in 

Prehistoric Greece', in History of the Hellenic World: 

Prehistory and Protohistory, University Park, Penn. (1974): 


, Peuples Prehellemques d'Origine Indo-Europeenne, 

Athens, 1977. 

Van Royen, R. A. and B. H. Isaac, The Arrival of the 
Greeks: The Evidence from the Settlements, Amsterdam, 


Danov, C, Althrakien, Berlin and New York, 1976. 
Detschew, D.., Die Thrakische Sprachreste, Vienna, 1957. 
Hoddinott, R. F., The Thracians, London, 1981. 
Polome, E., 'Balkan Languages', CAH 3/1 (1982): 866-888. 
Vulpe, R. (ed.) Actes du He Congres International de 

Thracologie, 3 vols, Bucharest, 1980. 

Katicic, R., Ancient Languages of the Balkans, The Hague- 
Paris, 1976. 

Prendi, F., 'The Prehistory of Albania' CAH 3/1 (1982): 


Stipcevic, A., The Illyrians, New Jersey, 1977. 

Baran, V. D. (ed.) Problemy Etnogeneza Slavyan, Kiev, 1978. 
Birnbaum, H., 'The Original Homeland of the Slavs and the 

Problem of Early Slavic Linguistic Contacts', JIES 1 

(«973): 407 421. 

, Common Slavic, Cambridge, Mass., 1975. 

Chropovsky, B. (ed.), Rapports du Hie Congres International 

d'Archeologie Slave, 2 vols, Bratislava, 1979. 
Gimbutas, M., The Slavs, London, 1971. 
Rybakov, B. A., Gerodotova Skifiya, Moscow, 1979. 
Sedov, V. V., Proiskhozhdenie i Rannyaya Istoriya Slavyan, 

Moscow, 1979. 

Trubachev, O., 'Linguistics and Ethnogenesis of the Slavs', 

JIES 13 (1985): 203-256. 
Werner, J., 'Zur Herkunft und Ausbreitung der Anten und 



Sklavenen', in Actes du Vile Congres International des 
Sciences Prehistoriques et Protohistoriques, Belgrad, 1971: 

Gimbutas, M., The Baits, London, 1963. 
Schmidt, W. P., Baltische Gewassernamen und das 

Vorgeschichtliche Europa, IF jj (1972): 1-18. 

Kriiger, B., Die Germanen, 2 vols, Berlin, 1983. 
Todd, M., The Northern Barbarians roo bc-adjoo, 

London, 1975. 

Barker, G., Landscape and Society, London, 1981. 
Durante, M. 'Lingua e Dialetti delPItalia Antica', in A. L. 

Prosdocimi (ed.), Popoli e Civilta dell Italia Antica, vol. 6, 

Padua, 1978. 
Pallottino, M., The Etruscans, London, 1975. 
Poultney, J. W., 'The Language of the North Picene 

Inscriptions', JIES 7 (1979): 49-64. 
Pulgram, E., The Tongues of Italy, Cambridge, Mass., 1958. 

, Italic, Latin, Italian, Heidelberg, 1978. 

Ridgway, D. and F. (eds), Italy before the Romans, London, 


Filip, J., Celtic Civilization and its Heritage, Prague, 1977. 
Greene, D., 'The Coming of the Celts: The Linguistic 

Viewpoint", Proc. VI th Internal. Congress of Celtic Studies, 

Dublin, (1983): 131-137. 
Harbison, P., 'The Coming of the Indo-Europeans to 

Ireland: An Archaeological Viewpoint', JIES 3 (1975): 


MacEoin, G., 'The Celticity of Celtic Ireland', in K. H. 

Schmidt (ed.), Geschichte und Kultur der Kelten, 

Heidelberg, 1986: 1 61-174. 
Mallory, J. P., 'The Origins of the Irish', Journal of Irish 

Archaeology 2 (1984): 65-69. 
Piggott, S., 'The Coming of the Celts: The Archaeological 

Argument', Proc. Vlth Interna t. Congress of Celtic Studies, 

Dublin (1983): 138-148. 
Powell, T. G., The Celts, London and New York, 1980. 
Savory, H. N., Spain and Portugal, London and New York, 


Schmidt, K. H., Die Festlandkeltischen Sprachen, 

Innsbrucker Beitrage zur Sprachwissenschaft 18, 1977. 
, Geschichte und Kultur der Kelten, Heidelberg, 1986. 
Untermann, J., Monumenta Linguarum Hispamcarum, 

Wiesbaden, 1975. 
Wagner, H., 'The Origin of the Celts in the Light of 

Linguistic Geography', Transactions of the Philological 

Society (1969): 203-250. 
Indo-European culture {general) 

Arntz, H. (ed.), Germanen und Indogermanen: V oik stum, 
Sprache, Heimat, Kultur; Festschrift fur Herman Hirt, 2 
vols, Heidelberg, 1936. 

Benveniste, E., Indo-European Language and Society, Coral 
Gables, Florida, 1973. 

Buck, C. D., A Dictionary of the Principal Indo-European 
Languages, Chicago, 1949. 

Cardona, G., H. Hoenigswald and A. Senn (eds), Indo- 
European and Indo-Europeans, Philadelphia, 1970. 

Carnoy, A., Les Indo-Europeens, Brussels, Paris, 1921. 

Crevatin, F., Richerche di Antichita Indeuropee, Trieste, 

Devoto, G., Origini Indeuropee, Florence, 1962. 
Feist, S., Kultur, Ausbreitung und Herkunft der 

Indogermanen, Berlin, 191 3. 
Gamkrelidze, T. V. and V. V. Ivanov, Indoevropeysky Yazyk 

i Indoevropeytsy, 2 vols, Tbilisi, 1984. 
Georgiev, V., Introduzione alia Storia delle Lingue 

Indeuropee, Rome, 1966. 
Hirt, H., Die Indogermanen, 2 vols, Strassburg, 1905-07. 
Mallory, J. P., 'Time Perspective and Proto-Indo-European 

Culture', World Archaeology 8 (1976): 44-56. 
Mayrhofer, M., W. Meid, B. Schlerath and R. Schmitt 

(eds), Antiquitates Indogermanicae, Innsbrucker Beitrage 

zur Sprachwissenschaft 12, 1974. 
Pokorny, J., Indogermamsches Etymologise hes Worterbuch, 

Bern, 1959. 

Polome, E. (ed.), The Indo-Europeans in the Fourth and 

Third Millennia, Ann Arbor, 1982. 
Scherer, A., 'Hauptprobleme der Indogermanischen 

Altertumskunde (seit 1940)', Kratylos 1 (1956): 3-21. 
, 'Indogermanische Altertumskunde (seit 1956)', 

Kratylos 10 (1965): 1-24. 
Schrader, O., Prehistoric Antiquities of the Aryan Peoples, 

London, 1890. 
Schrader, O. and A. Nehring, Reallextkon der 

Indogermanischen Altertumskunde, 2 vols, Berlin, Leipzig, 


Skomal, S. N. and E. Polome (eds), Proto-Indo-European: 
The Archaeology of a Linguistic Problem, Washington, 
D.C., 1987. 

Environment and Material Culture 

Adams, D. Q., 'Designations of the Cervidae in Proto-Indo- 
European', JIES 13 (1985): 269. 

Barber, E. J. W., 'The PIE Notion of Cloth and Clothing', 
JIES 3 (1975): 294-320. 

Diebold, R., 'Contribution to the Indo-European Salmon 
Problem', in W. Christie (ed.), Current Progress in 
Historical Linguistics, Amsterdam, 1976: 341-387. 

, The Evolution of Indo-European Nomenclature for 

Salmonid Fish: The Case of'Huchen, Washington, D.C., 

Friedrich, P., Proto-Indo-European Trees, Chicago, 1970. 

Hamp, E. P., 'Fish', JIES 1 (1973): 507-511. s 

Krogmann, W., 'Das Buchenargument', KZ 72 (1955): 1-29; 

73 (1956): 125. 
— , 'Das Lachsargument KZ 76 (i960): 161-178. 
Maher, J. P., '*H a e£mon: "(stone) axe" and "sky" in I-E/ 

Battle-axe Culture', JIES 1 (1973): 441-462. 
Mallory J. P., 'Indo-European and Kurgan Fauna I: Wild 

Mammals', JIES 10 (1982): 193-222. 
, 'Indo-European and Kurgan Fauna II: Fish', JIES 11 

(1983): 263-279. 
Kinship and Social System 

Beekes, R., 'Uncle and Nephew', JIES 4 (1976): 43-63. 
Bremmer, J., 'Avunculate and Fosterage', JIES 4 (1976): 

Friedrich, P., 'Proto-Indo-European Kinship', Ethnology 5 
(1966): 1-36. 

Gates, H., 'The Kinship Terminology of Homeric Greek', 
International Journal of American Linguistics, Memoir 27, 

Huld, M., 'CuChulainn and his IE kin', Zettschrift fur 

Celtische Philologie 38 (1981): 238-241. 
Scharfe, H., 'The Vedic Word for "King"', Journal of the 

American Oriental Society 105 (1985): 543 548. 



Sihier, A., The Etymology of PIE *reg- "king" etc.\ JIES 

5 (1977): 221—246. 
Szemerenyi, O., 'Studies in the Kinship Terminology of the 

Indo-European Languages', Acta Iranica 7 (1977): 1-240. 
Wikander, S., Der Arische Mdnnerbund, Lund, 1938. 
Wordig, F., 'A Generative-Extensionist Analysis of the 

Proto-Indo-European Kinship System' (Ph.D. thesis, 

Xerox Microfilms), 1970. 

Dumezil, G., 'L'Ari et les Aryas', Appendix II of Les Dieux 

Souverains des Indo-Europeens, Paris, 1977. 
Thieme, P., Der Fremdlmg im Rgveda, Heidelberg, 1938. 
Szemerenyi, O., 'Studies in the Kinship Terminology of the 

Indo-European Languages', Acta Iranica 7 (1977): 1-240 


Dumezil, G., 'Dieux Cassites et Dieux Vediques a Propos 
d'un Bronze du Louristan', Revue Hittite et Asianique 1 1 
(1950): 18-37. 

, V Ideologic Tripartte des Indo-Europeens, Brussels, 


, Les Dieux Souverains des Indo-Europeens, Paris, 1977. 

Grottanelli, C, 'Yoked Horses, Twins, and the Powerful 

Lady', J7£S 14 (1986): 125-152 
Larson, G. J., C. S. Littleton and J. Puhvel (eds), Myth in 

Indo-European Antiquity, Berkeley and Los Angeles, 1974. 
Lincoln, B., 'The Hellhound', 7 (1979): 273-285. 
, Priests, Warriors and Cattle, Berkeley and Los 

Angeles, 1981. 

Littleton, C. S., The New Comparative Mythology \ Berkeley 

and Los Angeles, 1982. 
Mallory, J. P., 'The Ritual Treatment of the Horse in the 

Early Kurgan Tradition', JIES 9 (1981): 205-226. 
O'Brien, S., 'Indo-European Eschatology: A Model', JIES 4 

(1976): 296-320. 
Puhvel, J. (ed.), Myth and Law Among the Indo-Europeans, 

Berkeley and Los Angeles, 1970. 
Puhvel, ]., 'Aspects of Equine Functionality', in Myth and 

Law, 1970: 159-72. 
, 'Remus et frater', History of Religions 15 (1975): 


, 'Victimal Hierarchies in Indo-European Animal 

Sacrifice, American Journal of Philology 99 (1978): 


Adrados, F. R., Die Rdumliche und Zeitliche Differ enzierung 
des Indoeuropdischen im Lichte der Vor- und Fruhgeschichte, 
Innsbrucker Beitrage zur Sprachwissenschaft Nr. 27, 

Barker, G., Prehistoric Farming in Europe ; Cambridge, 1985. 
Bosch-Gimpera, P., Les Indo-europeens: Problemes 

Archeologiques, Paris, 1961. 
Childe, V. G., The Aryans, London, 1926. 
Crossland, R. A., 'Immigrants from the North, CAH 1/2 

(1971): 824-876. 
Danilenko, V. N., Eneoltt Ukratny, Kiev, 1974. 
Diakonov, I. M., 'On the Original Home of the Speakers of 

Indo-European', JIES 13 (1985): 92-174. 
Dressier, W., 'Methodische Vorfragen bei der Bestimmung 

der "Urheimat"', Die Sprache 11 (1965): 25-60. 
Gamkrelidze, T. and V. Ivanov, 'The Ancient Near East 

and the Indo-European Question [and] The Migration of 

Tribes Speaking Indo-European Dialects, JIES 13 (1985): 


Gimbutas, M., 'The Beginning of the Bronze Age in Europe 
and the Indo-Europeans: 3500-2500 bc', JIES 1 (1973): 

, 'The First Wave of Eurasian Steppe Pastoralists into 

Copper Age Europe', JIES 5 (1977): 277"338- 
, 'The Kurgan Wave 2 (c. 3400-3200 bc) into Europe 

and the Following Transformation of Culture', JIES 8 

(1980): 273-315. 
, 'Primary and Secondary Homeland of the Indo- 
Europeans', JIES 13 (1985): 185-202. 
Gornung, B., K Voprosu ob Obrazovaniy Indoevropeyskoy 

Obshchnosti, Moscow, 1964. 
Koppers, W. (ed.), Indogermanen- und Germanenfrage; Neue 

Wege zu ihrer Losung, Salzburg and Leipzig, 1936. 
Lamb, S., 'Linguistic Diversification and Extinction in 

North America', XXXV Congreso Internacional de 

Amertcanistas, 2 (1964): 457-464. 
Mallory, J. P., 'A Short History of the Indo-European 

Problem', JIES 1 (1973): 21-65. 
Mallory, J. P. and M. E. Huld, 'Proto-Indo-European 

"silver"', KZ 97 (1984): 1-12. 
Meid, W., 'Probleme der Ra urn lichen und Zeitlichen 

Gliederung des Indo-Germanischen', in H. Rix (ed.), 

Flexion und Wortbildung, Wiesbaden, 1976: 204-218. 
Meillet, A., Les Dialects Indo-Europeens, Paris, 1922. 
Pisani, V., Indogermanisch und Europa, Munich, 1974. 
Porzig, W., Die Gliederung des Indogermanischen 

Sprachgebiets, Heidelberg, 1954. 
Pulgram, E., 'On Prehistoric Linguistic Expansion', in M. 

Halle et al. (eds), For Roman Jakobson, The Hague, 1956: 


Renfrew, C, 'Problems in the General Correlation of 
Archaeological and Linguistic Strata in Prehistoric 
Greece: The Model of Autochthonous Origin', in R. A. 
Crossland and A. Birchall (eds), Bronze Age Migrations in 
the Aegean, London, 1973: 263-276. 

, Archaeology and Language: The Puzzle of Indo- 
European Origins, London, 1987. 

Scherer, A. (ed,)^ Die Urheimat der Indogermanen, 
Darmstadt, 1968. 

Schlerath, B., Die Indogermanen. Das Problem der Expansion 
eines Volkes im Lichte seiner Sozialen Struktur, 
Innsbrucker Beitrage zur Sprachwissenschaft 8, 1973. 

Schmid, W. P., 'Baltische Gewassernamen und das 
Vorgeschichtliche Europa', IF 77 (1972): 1-18. 

Thieme, P., Die Heimat der Indogermanischen Grundsprache, 
Wiesbaden, 1954. 


Burrow, T., The Sanskrit Language, London, 1955. 
Collinder, B., 'Hat das Uralische Verwandte?', Acta 

Universitatis Upsaliensis 1 (1965): 4. 
Hajdu, P., 'Uber die Alten Siedlungsraume der Uralischen 

Sprachfamilie', Acta Linguistica 14 (1964): 47^83. 
, 'Finnougrische Urheimatforschung', Ural- A It disc he 

Jahrbucher 41 (1969): 252-264. 
Joki, A. J., Uralier und Indogermanen, Helsinki, 1973. 
Toivonen, Y., 'Zur Frage der Finnisch-Ugrischen 

Urheimat', Suomalais-Ugrilaisen Seuran Aikakauskirja 56 

(1952): 1-41. 

Uesson, A-M., On Linguistic Affinity: The Indo-Urahc 

Problem, Malmo, 1970. 
Vuorela, T., The Ftnno-Ugric Peoples, Bloomington, 1962. 

Bomhard, A., 'The "Indo-European-Semitic" Hypothesis 



Re-examined', JIES 5 (1977): 55-99. 
Brunner, L., Die Gemeinsamen Wurzeln des Semitischen 
und Indogermanischen Wortschatzes, Bern and Munich, 

Diakonov, I. M., 'On the Original Home of the Speakers of 

Indo-European', JflES 13 (1985): 92-174. 
Gamkrelidze, T. and V. Ivanov, 'The Ancient Near East 

and the Indo-European Question [and] The Migration of 

Tribes Speaking Indo-European Dialects', JIES 13 

(1985)- 3^9i- 
Hodge, C. T., 'Indo-Europeans in the Near East', 

Anthropological Linguistics 23 (198 1): 227-244. 
Levin, S., The Indo-European and Semitic Languages, 

Albany, New York, 197 1. 

Gamkrelidze, T. and V. Ivanov, "Kartvelian and Indo- 
European: A Typological Comparison of Reconstructed 
Linguistic Systems', in To Honor Roman Jakobson, I 
(1967): 707-717. 

, 'The Ancient Near East and the Indo-European 

Question [and] The Migration of Tribes Speaking Indo- 
European Dialects', JIES 13 (1985): 3-91. 

Machavariani, G. I., 'On the Problem of Indo-European- 
Kartvelian (South Caucasian) Typological Parallels', in 
Trudy VII Mezhdunarodnogo Kongressa 
Antropologicheskhikh i Etnographicheskikh Nauk, Moscow, 
5, 1964: 658-661. 

Wagner, H., 'The Origin of the Celts in the Light of 
Linguistic Geography, Transactions of the Philological 
Society (1969): 203-250. 


Agapov, S. A., I. B. Vasiliev and V. I. Pestrikova, 

Khvalynsky mogilnik i ego mesto v eneolite Vostochnoy 
Evropy, in Arkheologiya Vostochno-Evropeyskoy Lesostepi, 
Voronezh, 1979: 36-63. 

Alekseevka, I. L., Zhensky obraz v antropomorfnoy 
skulpture epokhi eneolita, in Pamyatniki Drevnego 
Iskusstva Severo-Zapadnogo Prichernomorya, Kiev, 1986: 
43 _ 5°- 

Anthony, D. W., 'The "Kurgan Culture": A 

Reconsideration', Current Anthropology 27 (1986): 

Artemenko, I. I. et al. (eds), Problemy Epokhy Bronzy Yuga 

Vostochnoy Evropy \ Donetsk, 1979. 

, Arkheologiya Ukrainskoy SSSR, 3 vols, Kiev, 1985. 

Basin, S. G et al. (eds), Drevnyaya Istoriya Povolzhya, 

Kuibyshev, 1979. 
, Volgo-Uralskaya Step i Lesostep v Epokhu Rannego 

Met alia , Kuybyshev, 1982. 
Chernyakov, I. T. and N. M. Shmagliy, 'Derevyannye psalii 

yamnoy kultury', Arkheologiya 42 (1983): 10-16. 
Chernykh, E. N., 'Metallurgical Provinces of the 5th-2nd 

Millennia in Eastern Europe in Relation to the Process of 

Indo-Europeanization', JIES 8 (1980): 317-336. 
Formozov, A. A., Kamenny Vek i Eneolit Prikubanya, 

Moscow, 1965. 
, Problemy Etnokulturnoy Istorii Kammenogo Veka na 

Temtorit Evropeyskoy Chasty SSSR, Moscow, 1977. 
Futoryansky, L. I. et al. (eds), Problemy Epokhi Eneolita 

Stepnoy t Lesostepnoy Polosy Vostochnoy Evropy, 

Orenburg, 1980. 
Hausler, A., 'Anthropomorphe Stelen des Eneolithikums im 

Nordpontischen Raum', Wiss. Zeitschnft d. Martin- Luther 

Unwersitdt, Halle-Wtttenburg 15 (1966): 29-73. 

, Die Grdber der Alteren Ockergrabkultur zwischen Ural 

und Dnepr, Berlin, 1974. 
, Die Grdber der Alteren Ockergrabkultur zwischen Dnepr 

und Karpaten, Berlin, 1976. 
, 'Zur Altesten Geschichte von Rad und Wagen im 

Nordpontischen Raum', Ethnographisch-Archaeologische 

Zeitschnft 22 (198 1): 581-647. 
Krylova, L. P., 'Kernosovsky Idol (Stela)', in Eneolit i 

Bronzovy Vek Ukrainy, Kiev, 1976: 36-45. 
Kuzmina, E. E., 'Kolesny transport i problema etnicheskoy i 

sotsialnoy istorii drevnego naseleniya Yuzhnorusskikh 

Stepey', Vestnik Drevney Istorii, 4 (1974): 68-^7. 
, 'Rasprostranenie konevodstva i kulta konya u 

Iranoyazychnykh piemen Sredney Azii i drugikh narodov 

Starovogo Sveta', in Srednyaya Aziya v Drevnosti i 

Srednevekove, Moscow, 1977: 28-52. 
Lagodovska, O. F., O. G. Shaposhnikova and M. L. 

Makarevych, Mykhaylh ske Poselennya, Kiev, 1962. 
Mallory, J. P., 'The Chronology of the Early Kurgan 

Tradition (Part 1), JIES 4 (1976): 257-294 
, 'The Chronology of the Early Kurgan Tradition (Part 

2), JIES 5 (1977): 339-368. 
, 'The Ritual Treatment of the Horse in the Early 

Kurgan Tradition', JIES 9 (1981): 205-226. 
Masson, V. M. and N. Ya. Merpert et al, Eneolit SSSR, 

Moscow, 1982. 

Matyushin, G. Yu., Eneolit Yuzhnogo Urala, Moscow, 1982. 
Medvedev, E. I. et al. (eds), Problemy Arkheologti Povolzhya 

i Priuralya, Kuibyshev, 1976. 
Merpert, N. Ya., Drevneyshie Skotovody Volzhsko-Uralskogo 

Mezhdurechya, Moscow, 1974. 
, 'O plemenykh soyuzakh drevneyshikh skotovodov 

Vostochnoy Evropy', in Problemy Sovetskoy Arkheologii, 

Moscow, 1978: 55-63. 
Potekhina, I. D., 'O nositelyakh kultury Sredny Stog II po 

antropologicheskim dannym', SA 1 (1983): 144-154. 
Petrenko, A. G., Drevnee t Srednevekovoe Zhivotnovodstvo 

Srednego Povolzhya i Preduralya, Moscow, 1984. 
Saygin, N. I. et al. (eds), Problemy Epokhy Neolita Stepnoy 1 

Lesostepnoy Zony Vostochnoy Evropy, Orenburg, 1986. 
Shilov, V. P., 'Modeli skotovodcheshikh khozyaystv 

stepnykh oblasty Evrazii v epokhu Eneolita i rannego 

Bronzovoga Veka', SA, 1 (1975): 5-16. 
Shilov, Yu. A., 'Ostatki vozov v kurganakh yamnoy kultury 

Nizhnego Podneprovya', Arkheologiya 17 (1975): 53-61. 
Telegin, D. Ya., 'Eneoliticheskie stely i pamyatniki 

nizhnemikhaylovskogo tipa', Arkheologiya 4 (1971): 3-17. 

, Seredno-stogivska Kultura Epokhi Midi, Kiev, 1973. 

, 'O neoliticheskikh pamyatnikakh Podonya i Stepnogo 

Povolzhya', Arkheologiya 36 (198 1): 3-19. 
— — , Mezolitychni Pamyatky Ukrainy, Kiev, 1982. 

, Dereivka, BAR International Series 287, Oxford, 1986. 

Telegin, D. Ya. and I Potekhnia, Neolithic Cemeteries and 

Populations of the Dnieper Basin, Oxford, 1987. 
Vasiliev, LB., Eneolit Povolzhya, Kuybishev, 198 1. 
Zvelebil, M. (ed.), Hunters in Transition, Cambridge, 1986. 

Badetskaya, E. B., Arkheologicheskie Pamyatniki v Stepyakh 

Srednego Eniseya, Leningrad, 1986. 
Chlenova, N. L., 'Arkheologicheskie materialy k voprosu ob 

irantsakh doskifskoy epokhi i indoirantsakh', SA 1 (1984): 


Gcning, V., 'The Cemetry at Sintashta and the Early Indo- 



Iranian Peoples', JIES 7 (1979): 1-29. 
Gryaznov, M. P., The Ancient Civilization of Southern 

Siberia, New York, 1969. 
Khlobystina, M. D., 'Drevneyshie mogilniki Gornogo 

Altaya\ SA 1 (1975): 17 34- 
Kiselev, S. V., Drevnyaya Istoriya Yuzhnoy Sibiri, Moscow, 


Kuzmina, E. E., Drevneyshie Skotovody ot Urala do Tyan- 

Shanya, Frunze, 1986. 
Potemkhina, T. M., Bronzovy Vek Lesostepnogo Pritobolya, 

Moscow, 1985. 
Vinogradov, A., M. A. Itina and L. T. Yablonsky, 

Drevneyshee Naselenie Nizovy Amudari, Moscow, 1986. 

Gimbutas, M., 'The Beginning of the Bronze Age in Europe 
and the Indo-Europeans: 3500-2500 bc\ JIES 1 (1973): 

Kushnareva, K. Kh. and T. N. Chubinishvili, Drevnie 
Kultury Yuzhnogo Kavkaza, Leningrad, 1970. 

Munchaev, R. M., Kavkaz na Zare Bronzovogo Veka, 
Moscow, 1975. 

Southeast Europe 

Chernyakov, I. T. and G. N. Toshchev, 'Kulturno- 
khronologicheskie osobennosti kurgannykh pogrebeniy 
epokhi Bronzy nizhnego Dunaya', in Novye Materialy po 
Arkheologii Severo-zapadnogo Prichernomorya, Kiev, 1985: 

Dinu, M., 'Quelques Considerations sur la Periode de 

Transition du Neolithique a L'Age du Bronze sur le 

Territoire de la Moldavie\ Dacia 12 (1968): 129-139. 
Dodd-Opritescu, A., 'Les £lements "Steppiques" dans 

L'Eneolithique de Transylvanie\ Dacia 22 (1978); 87-97. 
Ecsedy, I., The People of the Pit-grave Kurgans in Eastern 

Hungary, Budapest, 1979. 
Garasanin, M., "The Eneolithic Period in the Central Balkan 

Area', CAM 3/1 (1982): 136-162. 
Georgiev, G., N. Ya. Merpert et al., Ezero: Rannobronzovoto 

Selishche, Sofia, 1979. 
Jovanovic, B., 'Some Elements of the Steppe Cultures in 

Yugoslavia', JIES 11 (1983): 3I43- 
Kalicz, N., Die Fruhbronzezeit in Nordost-Ungarn, Budapest, 


Marinescu-Bilcu, S. et al., 'Contributions to the Ecology of 
Pre- and Proto-Historic Habitations at Tirpesti', Dacia 25 
(1981): 7-31- 

Merpert, N. Ya., 'Iz istorii drevnev amnykh piemen' in N. L. 

Chlenova (ed.) Problemy Arkheologii Evrazii i Severnoy 

Ameriki, Moscow, 1977: 68-80. 
Morintz, S. and P. Roman, 'Aspekte des Ausgangs des 

Aneolithikums und der Cbergangsstufe zur Bronzezeit im 

Raum der Niederdonau', Dacia 12 (1968): 45-128. 
Necrasov, O., 'Physical Anthropological Characteristics of 

Skeletons from the Kurgan Graves in Romania', JIES 8 

(1980): 337-343 
Nestor, L and E. Zaharia, 'Sur la Periode de Transition du 

Neolithique a L'Age du Bronze dans L'Aire des 

Civilisations de Cucuteni et de Gumelnita', Dacia 12 

(1968): 17-43. 

Roman, P., 'Strukturanderungen des Endaneolithikums im 
Donau-Karpaten-Raum', Dacia 15 (1971): 31-169. 

Telegin, D. Ya., 'Ober Kulturelle Kontakte zwischen der 
Neo-aneolithischen Bevolkerung des Nordpontischen 
Gebiets und der Balkan-Donauregion', in Hugelbestattung 
in der Karpaten-Donau-Balkan-Zone wdhrend der 

Aneolithischen Periode, Belgrade, 1986: 37-44. 
Yarovoy, E. V., Drevneyshie Skotovodcheskie Plemena Yugo- 

Zapada SSR, Kishinev, 1985. 
Zbenovich, V. G., Pozdnetripolskie Plemena Severnogo 

Prichernomorya, Kiev, 1974. 
, 'K probleme svyazey Tripolya s eneoliticheskimi 

kulturami Severnogo Prichernomorya', in Eneolit i 

Bronzovy Vek Ukrainy, Kiev, 1976: 57-68. 
Northern and Central Europe 

Bader, O. N., D. A. Kraynov and M. F. Kosarev (eds), 

Epokha Bronzy Lesnoy Polosy SSSR, Moscow, 1987. 
Buchvaldek, M., 'Corded Pottery Complex in Central 

Europe', JIES 8 (1980): 393-406. 
, 'Zum Gemeineuropaischen Horizont der 

Schnurkeramik', Praehistorische Zeitschrift 61 (1986): 


, 'Die Mitteleuropaische Schnurkeramik und das 

Nordliche Schwarzmeergebiet', Pamatky Arkheologicke 77 
Pt. 2 (1986): 486-497. 

Hausler, A., 'Zu den Beziehungen zwischen dem 

Nordpontischen Gebiet, Sudost- und Mitteleuropa im 
Neolithikum und in der fruhen Bronzezeit und ihre 
Bedeutung fur das Indoeuropaische Problem', Przeglad 
Archeologiczny 29 (198 1): 101-149. 

, 'Der Ursprung der Schnurkeramik nach Aussage der 

Grab- und Bestattungssitten', Jahresschrift fur 
Mttteldeutsche Vorgeschichte 66 (1983): 9-30. 

, 'Kulturbeziehungen zwischen Ost- und Mitteleuropa 

in Neolithikum?', Jahresschrift fur Mitteldeutsche 
Vorgeschichte 68 (1985): 21-74. 

Kilian, L., Zum Ursprung der Indogermanen, Bonn, 1983. 

Kosko, A., 'On the Research of the Beginnings of the 
Parallel Developmental Ties of the Cultural Systems of 
the Boundary of Eastern and Western Europe', Memoires 
Archeologiques (Lublin) (1985): 37-49. 

Krzak, Z., 'Der Ursprung der Schnurkeramischen Kultur', 
Germania 59 (1981): 21-29. 

Lichardus, J., Rbssen-Gatersleben-Baalberge, Bonn, 1976. 

— , 'Zur Funktion der Geweihspitzen des Types Ostorf,' 
Germania 58 (1980): 1-24. 

Menk, R., *A Synopsis of the Physical Anthropology of the 
Corded Ware Complex as the Background of the 
Expansion of the Kurgan Culture, JIES 8 (1980): 361- 

Merpert, N. Ya., 'Drevneyamnaya kulturno-istoricheskaya 
oblast i voprosy formirovaniya kultur shnurovoy keramiki', 
in L. V. Koltsov et al. (eds), Vostochnoy Evropa v Epokhu 
Kamnya i Bronzy, Moscow, 1976: 103-127. 

Neustupny, E., 'Economy of the Corded Ware Cultures', 
Archeologicke Rozhledy 21 (1969): 43-67. 

Schwidetzky, I., 'The Influence of the Steppe People Based 
on the Physical Anthropological Data in Special 
Connection to the Corded Ware-Battle Axe Culture', 
JIES 8 (1980): 345-360. 

Sulimirski, T., Corded Ware and Globular Amphorae 
Northeast of the Carpathians, London, 1968. 

The process of expansion 

Barth, F., Process and Form in Social Life, London, Boston, 

, Features of Person and Soaety in Swat, London, 

Boston, 1 98 1. 

Fasold, R., The Sociolinguistics of Society, Oxford, 1984. 
Khazanov, A. M., Nomads and the Outside World, 
Cambridge, 1984. 



Lefebure, C. (ed.), Pastoral Production and Society, 
Cambridge, 1979. 

Sherratt, A., 'Plough and Pastoralism: Aspects of the 
Secondary Products Revolution', in I. Hodder, G. Isaac 
and N. Hammond (eds), Pattern of the Past: Studies in 
Honour of David Clarke, Cambridge, England, 1981: 261- 


Krader, L., 'Buryat Religion and Society*, Southwestern 
Journal of Anthropology 10 (1954): 322-351. 

Littleton, C. S., The New Comparative Mythology, Berkeley 
and Los Angeles, 1982. 

Mallory, J. P., 4 A Short History of the Indo-European 
Problem', JflES 1 (1973): 21-65. 

Poliakov, L., The Aryan Myth, London, 1974. 

Sources of illustrations 

Monochrome plates 

1, 2 Hirmir Fotoarchiv. 3 Ankara Museum. Photo J. Powell. 4 Cincin- 
nati Art Museum, No. 1057.29. 5 Original in State Hermitage, Lenin- 
grad. From electrotype in Victoria and Albert Museum, London. Photo 
P. Clayton. 6 Chester Museum. 7 British Museum, London. 8, 9 Musee 
Guimet, Paris. Photos Giraudon. 10 National Museum, Athens. Photo 
Hirmir. 11 National Museum, Athens. Photo J. Powell. 12 Archaeol- 
ogical Museum, Lovech. 13 A. Kastelicand the Archaeological Museum, 
Ljubljana. 14 Ashmolean Museum, Oxford. 15 Courtesy the Arch- 
aeological Institute of the Slovakian Academy of Sciences, Nitra. 
16 Muzeum Archeolog znego w Poznaniu. 17 National Museum of 
Antiquities, Copenhagen. 18 Statens Historiska Museum, Stockholm. 
19 Museo Civico, Bologna. From Kastelic, Situla Art. 20 National 
Museum, Copenhagen. 21, 23 Photos courtesy I. B. Vasiliev. 22, 

26 Photos courtesy D. Telegin. 24, 25 Photos courtesy I. Potekhina. 

27 Drawing by E. Brennan. 28 Magyar Nemzeti Muzeum, Budapest. 

Line illustrations 

Unless otherwise credited, line illustrations are courtesy the author. For 
full details of books see Bibliography. 

Quotation (p. 9): J. Parsons, The Remains of Japhet, London, 1767: iii. 
2 After G. Rohlfs, Romanische Sprachgeographie, Munich, 197 1: 279. 3, 
4 After J. Parsons, The Remains of Japhet, London, 1767: 316-317; 340. 
Abridged and revised. 5 After E. Ben v enisle, 1973: 530-531, and revised, 
with permission of Faber and Faber Ltd. 6 W. P. Lehmann and L. 
Zgusta, 'Schleicher's Tale After a Century', in B. Brogyanyi (ed.), Studies 
in Diachronic, Synchronic and Typological Linguistics: Festschrift for 
Oswald Szemerenyi, Amsterdam, 1979: 455-466. 7 After H. Pedersen, 
193 1 : 312. 8 After O. Schrader, 1890: 65. 9 After F. Adrados, Die 
Rdumliche und Zeitliche Differenizierung des Indoeuropdischen im Lichte 
der Vor- und Fruhgeschichte ( = Innsbrucker Beitrage zur Sprachwissen- 
schaft 27), 1982: 25, and slightly abridged. 10 After T. Gamkrelidze and 
Vy. Ivanov, 1984 vol. 1: 415. 11 After R. Anttila, 1972: 305, with 
permission of the author. 12 Excerpts from James Gordon, The English 
Language, New York, 1972: 157 -158. Quotation (p. 24): Hans von 
Wolzogen, 1875. 'Der Ursitz der Indo-Germanen', Zeitschrift fur 
Vblkerpsycholgie 8, 1875: 2. 15, 16 Hattic text and Hurrian vocabulary 
from Friedrich J. et al., 1969: 458; 12, 16. 19 Texts from O. Haas, 1966: 
66, 189. 20 Text and translation from W. B. Lock wood, 1972: 180. 

21 After A. Jackson, An Avestan Grammar, Stuttgart, 1892: xxxi-xxxii. 

22 After A. Meillet and M. Cohen, Les Langues du Monde, Paris, 1952: 
map XIa. 25 From German translation of A. Kammenhuber, Hippologta 
Hethitica, Wiesbaden, 1961: 54-55. 28 Adapted from S. P. Gupta, 1979: 
241.29 After B. and R. Allchin, 1982: 238. 31 After R. Kent, Old Persian, 
1950: 146, with permission of the American Oriental Society, New 
Haven, Ct, USA. 32 After M. Winn, 1974: i27.33Textto Yasna46from 
H. Humbach, Die Gathas des Zarathustra, Heidelberg, 1959: 128; 
translation from J. Duchesne-Guillemin. The Hymns of Zarathustra, 
London, 1952: 75, with permission of John Murray Ltd. 34-35 After A. 
Mandelshtam, 1968: 15, 27. 36 Text from W. Thomas, 'Ein Tochar- 
ischerLiebesbrief, KZ71, 1954:78-80.39 Location of languages after I. 
Diakonov, 1968. Quotation (p. 66): W. Ripley, The Races of Europe, 
London, 1900: 453. 40 After J. T. Hooker, Linear B: An Introduction, 
Bristol, 1980: 104-105. 41 Translation from J. T. Hooker. 1976: 12, with 
permission of Routledge and Kegan Paul Ltd. 42 Adapted from N. G. L. 
Hammond, 1976: 132. 43 After S. Hiller, 'Zur Frage . . .', 1982: tafel 1. 
44 Poem by D. Agolli in M. Gerecaliu, Q, Haxhihasani and J. Panajoti 
(eds), Ceshtje te Folklorit Shqiptar I, Tirana 1982: 358. 45 Map after A. 

Stipcevic, 1977: 32; quote from Appians Roman History, vol. 2, 55-57, 
translated by Horace White, Loeb Classical Library no. 3, with 
permission of Heinemann Ltd, 1 .ondon. 46 Texts from W. B. Lockwood, 
1972: 165-171. 48, 50 After V. Sedov, 1979: 20; 79. 49 After V. Baran, 
1978: 18. 51-53 After B. Rybakov, 1979: 201; 197; 207. 54 After M. 
Gimbutas, 1963: 30. 55 From S. Chatterji, Baltic and Aryan, Simla, 
1968: 159-160, with permission of the Indian Institute of Advanced 
Study. 58 Adapted from B. Kruger, 1983, vol. 1: 96-97, 384-385. 
59 After E. Pulgram, 1958: 198, with minor additions. 60 G. and L. 
Bonfante, The Etruscan Language, Manchester, 1983. 61 Translation by 
J. Poultney, The Bronze Tablets of Iguvium, 1959: 158, with the 
permission of the American Philological Association. 62 After E. 
Pulgram, 1978: 66-68, with permission of C. Winter Universitatsverlag; 
and M. E. Huld, pers. com. 63 Text from J. Poultney, 1979: 50. 64 From 
E. Pulgram, 1978: 53, with permission of C. Winter Verlag. 66 After M. 
Gimbutas, 1973: 195. 69 After M. Szabo, The Celtic Heritage in Hungary, 
Budapest, 1971: 8. 70 After J. Untermann, 1975: 108. Quotation (p. 
110): M. Haas, The Prehistory of Languages, The Hague, 1969: 32. 
73 Drawn from information in F. Wordig, 1970. Quotation (p. 128): P. 
Thieme, 'The Comparative Method for Reconstruction in Linguistics', 
in D. Hymes (ed ), Language in Culture and Society, New York, 1964: 593. 
74-^77 After R. Dussaud, 'Anciens Bronzes du Louristan et Cultes 
Iraniens', Syria 26 (1947): 213. Redrawn by E. Brennan. 78-^79 After B. 
Lincoln, 1981: 114, 160. Quotations (p. 143): A. Sayce, An Introduction 
to the Science of Language, London, 1880: 121 . A. Sayce, An Introduction 
to the Science of Language, London, 1890: 121. A. Sayce, 'The Aryan 
Problem - Fifty Years Later', Antiquity 1, 1927: 204-215. 81 After D. 
Pitcher, An Historical Geography of the Ottoman Empire, Leiden, 1972: 
maps I-IV. 82 See ref. ill. 22: map vii. 84 After P. Friedrich, 1970: 1 13, 
with addition. 85 After M. Gimbutas, 1977: 331, and amended by 
Gimbutas 1985. Quotation (p. 186): B. Lincoln, 1981: 181. 86-88 After 
I. Artemenko et al., 1985: vol. 1, 112, 136. 90, 92 After V. Masson, N. 
Merpert et al., 1982: 101; 287. 91, 93 After Artemenko et al., 1985: 204. 
94 After D. Telegin, Dereivka, Oxford, 1986: 7. 95 Courtesy of D. Ya. 
Telegin. 96-99 After D. Telegin, 1973: 138, 88, 73, m. 100-106 After 
Artemenko et al., 1985: 314, 314, 312, 282, 333, 326, 326. 107 After R. 
Munchaev, 1975: 150. 110-111 After I. Vasiliev, 198 1: 122-123. 
108 After I. Vasiliev and G. Matveeva, U Istokov Istorii Samarskogo 
Povolzhya, Kuybishev, 1986: 39. 112 After N. Merpert, 1974: 153. 
113 After O. Lagodovska, O. Shaposhnikova, and M. Markarevic. 1962: 
66. 114 After Yu. Shilov, 'Ostatki vozov . . .*, 1975: 54. 115 After I. 
Chernyakov and N . Shmagliy, 1983: 13. 11 6-1 18 After Artemenko etal., 
1985: 341, 343, 343. 119 After A. Hausler, 1966: 64. 120 After I. 
Alekseevka, 1986: 45. 121 After Hausler, 1966: 64. 122 See ref. ill. 108: 
39. Quotation (p. 222): R. von Ihering, The Evolution of the Aryan, 
London, 1897: 5. 123-124 After E.Badetskaya, 1986: 20,21. 126 Adapted 
from N. Chlenova, 1984: 100- 101. 128 After K. Kushnareva and T. 
Chubinishvili, 1970: 60-61. 129 After Gimbutas, 1977: 315. 131 After 
Masson and Merpert et al., 1982: 300. 132-134 After Artemenko et al., 
1985: 250. 136 After I. Chernyakov and G. Toshchev, 1985: 8. 137 After 
I. Ecsedy, 1979: 29. 138 Adapted from M. Buchvaldek and D. Koutecky, 
Vikletice, Prague, 1970: 105. 140 After O. Bader et al., 1987: 61. 
144 After T. Sulimirski, Prehistoric Russia, London, 1970: 163. 143 After 
Artemenko et al., 1985: 288. 145 After L. Kilian, 1983. Quotation (p. 
266): G. Vacherde Lapouge, L' Aryen, Son Role Social, Paris, 1899: 372- 
373» quoted in L. Poliakov, The Aryan Myth, London, 1974: 270. 
147 From H. Klein, Graphic Worlds of Peter Breugel the Elder, New 
York, 1963: 151, with kind permission of Dover Publications, New York. 



I would like to acknowledge the questionable honour bestowed upon me 
by the late Glyn Daniel who convinced me that I should tackle a subject 
wherein I would not only be able to run afoul of almost every regional 
specialist in archaeology from Ireland to China but also wander outside 
my own discipline and subject myself to the criticism of my colleagues in 
linguistics, mythology and anthropology. Except for a career in politics, 
one seldom gets the opportunity to offend so many people at once. 

Preparation of this book was greatly facilitated by the British Council 
and the Academies of Science of both the USSR and the Ukrainian SSR. 
I would like to thank especially Nikolai Merpert (Moscow) and Dmitry 
Telegin and all his associates in the Institute of Archaeology (Kiev) for 
their generous assistance and hospitality. In addition, I would like to 
thank Igor Vasiliev (Kuibyshev) for reading a portion of the manuscript 
and supplying me with several illustrations; Alexander Hausler (Berlin- 
Halle) for over a decade of book and article exchanges which has made it 
possible for me to obtain numerous works I would otherwise have been 
unable to consult. The production of maps and figures was greatly 
facilitated by the assistance of Paul Campbell and Evelyn Cooper of 
December Publications, Belfast, and Emma Brown, Queen's University, 
who prepared most of the illustrations. I would thank Louise Porter of the 

Queen's secretarial centre who endured several drafts of the text, Barrie 
Hartwell of the Archaeology Department for photographic assistance, R. 
B. Warner of the Ulster Museum who permitted me access to his 
programme for graphically representing radiocarbon dates, J. B. Rutter 
(Dartmouth) who provided valuable comments on my treatment of 
Greek origins, and Stuart Piggott who read through the entire text. I owe 
a special debt to M. E. Huld (California State University, Los Angeles) 
who read the first eight chapters making many valuable comments and 
preserving me from numerous linguistic errors. None of these is 
responsible for any infelicities of fact or interpretation that I may have 
stubbornly or unwittingly persisted in, nor are any of those who originally 
instructed me in the languages, mythology and archaeology of the Indo- 
Europeans, among whom I would mention Raimo Antilla, Ken Chap- 
man, P. K. Ford, Hans-Peter Schmidt, C. Scott Littleton,JaanPuhvel,J. 
Caerwyn Williams, Lilli Kaelas and most especially Marija Gimbutas - 
who if not entirely in agreement with my conclusions might at least draw 
some amusement that I have at last overcome my disdain for intuition. 
Finally, very special thanks to my wife Eimear for both her assistance and 
patience which saw me through to the end of this book. 


Numbers in italics refer to line drawings; 
numbers in bold refer to plates 

Aesir 139 

Afanasievo culture 62, 223-6, 262; j8, 123, 

124, *2S 
Afrikaans 84 
Akkadian 122, 150 
Alaca Hiiyiik 30, 168, 232 
Alans 48, 53, 61, 229 
Albanian 73-4, 92, 156; 3, 44 
Alexandria 198; 26 
Altaic 151, 159, 268, 270 
Altheim culture 217 
Altin Depe 52, 55, 231 
Anatolian 27-9, 34, 39, 64, 66, 1 13, 154-6, 

158, 182, 222, 231, 241, 263-4 
Andronovo culture 52-3, 61-2, 227-31; j8 y 

726, 127 
Anglo-Saxon 166, 168, 266 
Antes 76-7 
Apam Napat 129 
Aphrodite 132 

Aradvi Sura Anahita 133, 220 

Armenian 33-5, 58, 61, 63-4, 108, 152-5, 178, 

182, 222, 241, 263 -4, 273; ?, 20 
Arpa 62 

Artemenko, I. 247-8 
Aryaman 130 

Aryan 125-6, 138, 141, 143, 145, 258, 261, 

266-70, 276-7 
Asklepios 132 

Assyrian 24-5, 28, 31, 35, 49-50, 63 
asvamedha 135-7 
Asvins 119, 133, 135 
Athena 132 

Avesta 52, 54, 132-3, 229; 21,33 

Baalberge group 249 50, 252, 254; 141 
Bactrian 51, 53 

Baden culture 71, 109, 163, 218, 239, 250; 28 
Badetskaya, E. 225-6 
Balkan-Danubian complex 109; 72 

Baltic 18, 58, 77-8, 81-4, 108, 150, 153, 155, 
157, 182, 223, 243, 255-6; 5J, 34\ 16 

Balto-Slavic 18, 78 

Baran, V. 78 

Barker, G. 177, 260 

Barth, F. 259-60 

Basque 105, 145, 147, 181; 70 

battle-axe 70, 71, 93, in, 122, 200, 204, 213, 
219, 244, 246, 248; 138 

Beaker culture 168, 236 

beaver 116, 179, 187, 200, 212, 217 

beech 13, 112, 1 15-16, 160, 216, 268, 275; 84 

Bengali 36, 3 

Benveniste, E. 118-19, I2 5 

Berezhnovka 215 

Beycesultan 29 

Bhaga 130 

Bibikova, V. 221 

Biedenkapp, G. 143, 269 

birch 115, 117, 161, 187, 228, 275 

Bishkent culture 53 

Blegen, C. 68 

boar 135, 198, 202, 212, 226, 244; 96 
Bodrogkeresztur culture 250 
Bokonyi, S. 218, 273 
Bopp, F. 14 
Brahui 44 

Brandenstein, W. 269 
Breton 18, 96, 166 
Brittonic 96, 106, 166, 266 
Bryges 32 

Buchvaldek, M. 247-50, 264 

Buddhism 56, 58—60 

Bug-Dniester culture 189-190, 196-7 

Bulgarian 77; 46 

Burrow, T. 42-3, 52 

Byelorussian 77 

Caesar 85, 95, 141-2 
Carnoy, A. 270 
Castor 132, 135 
Catal Huyiik 178 
cattle 52, 111-12, 117, 

159, 183, 187, 189-91, 193, 199, 203, 
206-7, 2I 3 _1 4> 21 7> 220, 223, 228, 237, 
250, 275 

cattle cycle 137-8; 79 

cattle raid 1 17-18, 137-8; 78 

cattle sacrifice 118, 137-8 

Causasian languages 26, 30, 50, 151, 231 

Cayonu 178 

Celtic 10, 12, 18, 20, 58, 61, 63, 74, 89, 95-6, 
105-8, 113, 133, 1 41-2, 145, 147, 152-5, 
157-8, 165, 167, 182, 223, 243, 256, 264, 
270, 274; 67, 69, 70; 20 

Cemetery H culture 47; 28 

centum-satem 19, 58, 153-4; 8, 11 

Cernovoda I culture 238, 253 

Chapli 202; 101, 102 

chariot 38-42, 45, 53, 55, 69, 136-7, 228; 11 

Chernoles culture 81, 84; 52 

Chernyakov, I. 239 

Chernyakovo culture 79-80; 50 

Chernykh, E. 201 

Childe, V. G. 143, 259, 266, 270 

Chinese jo, 156, 267, 270; 4 

Chlenova, N. 230 

Chubinisvili, T. 233 

Comb-Pit Ware culture 149, 256 

copper 47, 93, 121, 164, 179, 184, 191, 195, 

200, 202-3, 2 °5> 2 °7, 2I 3 _1 4» 21 8, 223, 

226, 233, 235-6, 238, 240, 244; 1 01, 102 
Corded Ware culture 108-9, I22 > 182, 185, 

197, 223, 244-50, 252-7, 261, 264; 72, 138, 

/jo, 140 
Cornish 96 
Cofofeni culture 239 
Cris, culture 189, 194 
Croatian 77 
Crouwel, J. H. 41 
Csongrad-Kettoshalom 235 
Cucuteni culture (see Tripolye) 
Cucuteni C ware 235, 238 
Czech 77; 46 

133, 135, 138, 1 50-1, Dacian 72, 79, 182 



Dani, A. H. 48 
Danilenko, V. 248 
Danish 84; 3 
Darius 126; 31 
dasa 45, 141 
Decea Muresului 235 
Demerci Hiiyiik 29 

Dereivka 198-201, 217, 220-1; 94, 95, 06, 97, 

Diakonov, I. 34-5, 41, 51, 150, 181, 262, 274 
Diebold, R. 160 
Dius Fidius 140 

Dnieper-Donets culture 190- 1, 197, 201, 

205-7, 210, 219, 252-3, 256-7, 262; 86, 

no; 22, 24, 25 
dog in, 113, 119, 187, 191, 198-9, 203, 214, 

220, 228, 275 
Dolukhanov, P. 189 
Dravidian 44-5, 156, 159 
Dumezil, G. 130-4, 136-7, 139-142, 219, 

270-1; 74, 75, 76, 77 
Dumitrescu, V. 234 
Durkheim, E. 130 
Dutch 11, 84; j 
Dyaus 128; 19 

Ecsedy, I. 236, 241 

Elamite 44, 50, 63-4, 147 

Elamo-Dravidian 44-5; 27; 7 

English 19, 22-3, 84, 156, 159, 166, 259, 

265-6; 3, / 2, 36, 57 
Epona 119 
Este culture 91-2 
Esus 139 

Etruscan 84, 88-9, 94, 147, 181; 60 
Ezero 71, 184, 238-9, 241 

Faliscan 90 

Fatyanovo culture 249, 261 
Fedeleseni 234 
Feist, S. 269 

fertility deities (third function) 13 1-3, 135, 
139, 142 

Finnish 77, 84, 145, 148, 268, 270 
Finno-Ugric 147 -9, 152, 261-2 
fish 116, 189-191, 198, 200 
Fjorgyn 129; 55 
Formozov, A. 191, 256 
French 19, 89, 156, 265-6; 3 
Frey 132 
Freyr 132 

Friedrich, P. 113, 161 

Gaelic 166 
Gallehus horn 78 

Gamkrelidze, T. 7, 122, 141, 143, 150, 163, 

181-2, 216, 274; 10 
Gandhara Grave culture 24, 47, 48, 230-1; 28, 

29, 127 
Garasanin, M. 234 
Gathas 52; 33 
Gaudo culture 95 
Gaulish 95, 139 
Geiger, L. 268 
Genning, V. 228 
Georgiev, G. 239 
Georgiev, V. 274 
Gerasimovka 220 
German 84; j, 57 

Germanic 10, 14, 18-19, 58, 61, 77, 80-2, 
84-7, 9i > i<>7^8, i", "3, 124, 130, 140, 
MS, »50» 152-5, 157-8, 164, 182, 223, 243, 

256, 262, 264, 266-7, 270; 56, 57, 58; 17 
Getae 72, 79 

Ghirshman, R. 39-40, 42, 50, 51; 26 
Gimbutas, M. 79, 182-5, 215, 231-2, 234, 
241, 248, 250-1, 253, 256, 260, 262; 85 
Glasinac culture 76 
Glob, P. V. 248 

Globular Amphora culture 182, 243, 249-53, 

255-0; 142-4 
glottochronology 276 

goat in, 1 17-18, 135, 149, 191, 199, 203, 

207, 211, 214, 223, 228 
gold 47, 121, 149, 205, 223 
Gonda, J. 125 
Gordion 31, 33 
Gornung, B. 181 
Gothic 8, 12, 77, 79, 84; 50, 57 
Grantovsky, E. 54 

Greek 8, 10, 12, 25, 30, 34-5, 56, 58, 61, 
63-74, 84, 87, 89, 95, 107-8, 129, 153-6, 
161, 180-2, 222, 241, 261-4, 270, 272, 274; 
3, 40, 41, 42, 43; 10 

grey wares 39-40, 47~52, 54"5, 229; 32 

Grushevka 220 

Gumelnitsa culture 196-7, 233, 238 

Haas, M. no 
Habur Ware 39 

Hausler, A. 219, 247, 251-2, 254-5 

Haley, J. 68 

Hallstatt culture 106 

Hamito-Semitic 159 

Harpstedt culture 87; 5# 

Hattic 26, 30, 63, 64, 147, 178-9, 181, 232, 

273; '5 
Hattusa 25, 31, 33, 37, 131 
Haugen, E. 141 
Hawkes, C. 64 
Hebrew 10, 267; 4 
Heine-Geldern, R. 59 
Helen of Troy 132 
Hengist 135, 141 
Hera 132 

Herakles 130, 137; 5 

Herodotus 32, 33, 35, 72, 81, 83-4, 132; 41 
Hindi 36, 259 
Hirt, H. 16, 269; 6 
Hispano-Celtic 95, 105, 107; 70 
Hissar 39, 40, 51, 55 

Hittite 16, 25-6, 28-^, 30, 32 34-5, 37, 58, 
6t, 63, 133, 136, 150, 156, 168, 178, 231-2, 
273; '5, 25; 1, 2 

Horsa 135, 141 

horse 16-17, 29, 30, 38, 40-2, 45-8, 53, 55, 
59, 62, 69-70, 93, in, 119, 126-7, 13 1-3, 
135-7, 142, 158, 161-3, 179, 183-5, 187, 
191, 193, * 95, 198-9, 203-4, 206-7, 
210-12, 214, 217, 219-21, 223, 228, 230, 
232, 234, 237-8, 242, 244, 250, 252-4, 259, 

262, 270, 273; 122, I2Q 

Hsiung-nu 59-60 
Hulthen, B. 252 

Hungarian 11, 145, 148, 261, 270 
Huns 61, 223 
Hupasiya 137 

Hurrian 26, 30, 34, 37-42, 50, 55, 63-4, 147, 

150, 163, 178-9, 232-3; 16; 1 
Huxley, T. 269 
Hymir 137 

Iberian 95, 105, 147, 181; 70 
Icelandic 84 

Igren 201 
Iguvium 90; 61 
von Ihering, R. 222 
Iliad 67, 72, 139, 141; 11 
Illyrian 32, 72-5, 90, 108, 153, 182, 264; 4$; 
13, M 

Indo-Aryan 10, 18, 20, 35-47, 51-3, 55-6, 58, 
63, 66, 72, 108, in, 113, 133, 150, 152, 
155-6, 228-9, 231, 241, 256, 263, 270, 272, 
274; 21, 22, 23, 26; 8, 9 

lndo- Iranian 36, 38-40, 52-4, 61, 63, 78, 113, 
119, 125-6, 149, 152-4, 158, 178, 180, 182, 
185-6, 217, 223, 226-9, 231, 258, 262-3; 21 

Indo-Semitic 159 

Indo-Lralic 149, 159 

Indra 37, 42, in, 13 1-3; 23, 9 

Indus Civilization 43, 45-7, 228-9, 274; 27, 
'27; 7 

Indus script 44; 7 

Iranian 10, 18, 20, 33, 35-6, 38-40, 42-3, 48, 
50, 52-6, 58, 60-3, 77, 81, 108, 1 13-14, 
149, 152, 183, 185, 228-9, 231, 241, 259, 
261, 263, 270; 21, 30, 31, 32, 33, so; 4, 6 

Irish 10, 11, 18, 96, 106, in, 140, 178, 274; 

Italian 89; 3 

Italic 10, 18-20, 58, 61, 90, 107, 108, 153-5, 

180, 264; 59; 19 
Italo-Celtic 108 

Ivanov, V. 7, 122, 141, 143, 163, 1 81-2, 216, 
274; JO 

Jastorf culture 86, 164; 5<? 
Johanncsson, A. 269 
Jones, W. 11, 273 
Jupiter 128, 132, 140 

Kafiri 36, 47 

Kalanchak 204 

Kanes 25, 28-9 

Karanovo 73, 109, 233, 238 

Karashahr 56, 59, 60 

Kartvelian 26, 150-1, 159, 163, 231 

Kaskian 30 

Kassite 38-9 

Katicic, R. 74 

Kelteminar culture 226 

Kemi-Oba culture 197, 203-6, 262; 103, 106 

Kerberos 129 

Kernosovka 204, 220; 27 

Ketegyhaza 137 

Khlobystina, M. 219, 223, 225 

Khutor Khryashchevskogo 219 

Khvalynsk 197, 207, 210-11, 226; ///; 21 

Kikkuli 37, 41; 25 

Kilian, L. 246-7, 254-5; l 45 

Kimmerian 31, 48 

king in, 125, 141 

kinship 123-124, 179, 276; 73 

Kiselev, S. 226 

Klaproth, J. 268, 273 

Knossos 66, 68-9 

Kolochina culture 78; 49 

Komarov culture 8 1 ; 52 

Konstantinovka 198 

Koppers, W. 269 

Korucu Tepe 30 

Kosko, A. 25 1 

Kossinna, G. 164, 269 

Kozlowski, J. and S. 255 

Krader, L. 272 

Krahe, H. 276 



Kucha 57, 59-60 
Kuhn, A. 11 1 
Kurdish 48, 11 5- 16 

Kurgan tradition 182 5, 242-3, 254, 256, 262; 

Kuro-Araxes culture 29, 163, 232 3, 261; 107 
Kushnareva, K. 233 
Kuzmina, E. 219, 228-31, 277 

Lamb, S. 145-6 

de Lapouge, G. V. 266 

La Tene 105-6, 142, 164; 69 

Latham, R. G. 152, 268 

Latin 12, 19, 66, 73, 77, 82, 88-9, 94-5, 133, 

H7» 259; j, j6, 60, 6r, 64 
Latvian 81 

Lengycl culture 197, 253, 257 
Lepontic 95 
Levi-Strauss, C. 141 
Lichardus, J. 251-3 
Ligurian 89 

Lincoln, B. 137-8, 140-1; 186, 78 

Linear A 69, 180 

Linear B 66, 69, 274; 40 

Linear Ware culture 164, 178, 186, 189, 194, 

242, 253, 257, 277 
Lithuanian 81-2, 156-8; 55 
Littauer, M. A. 41 
Littleton, C. S. 271 

Lower Mikhaylovka 197, 203-6, 250-1; /oj, 
j 04, 105 

Luwian 25-7, 29, 30, 32, 34-5, 63, 68-9, 71, 

178, 231 
Lycian 25, 30 

McAlpin, D. 44, 45 
Macedonian (Greek) 74 
Macedonian (Slavic) 77 
Macha 135 

Mahabharata 46, 129, 139 
Maher, J. P. 122 
Maidanetskoe 236; rji 
Makkay, J. 260, 277 
Maliq 75, 109 
Mann us 140 
Manu 130, 140 
Manx 18 
Ma rat hi 36 
Marathon 69-70 
Mariupol 191, 206 
marriage 123, 141; 55 
Mars 132, 136 
Maruts in, 133 
Massagetae 53, 229 
Matwy group 251 
Matyushin, G. 192, 277 
Maykop 205, 232; 107 
mead 136, 275 
Medes 49, 54-5 

Merpert, N. 211, 215, 220, 239, 241, 254, 262 
Messapic 90-2, 108, 274; 62 
Midas 31, 33 

Middle Dnieper culture 248 
Mikhaylovka 203, 211-13, 217-18, 236, 241; 

Milograd culture 83 
Mingachaur 232 
Minoans 66 

Mitra/Mithra 37, 131-3, 140; 21 

Mitanni 37-43, 52, 55, 13 1-2, 150, 229; 24, 

25, 26 
Mohenjo Daro 45; 7 

Morintz, S. 260 
Morris, C. 269 
Muller, M. 267, 269 
Munchaev, R. 233 
Munda 44 
Muski 34-5 

Mycenae 66, 68-70; 10, n 

Nalchik 205, 244 
\amazga 52, 55, 231 
Narva culture 256 
Nasatya 37, 131 
Ncchtain 129 
Nehring, A. 269 
Neptunus 129 
Ncuri 83 

Neustupny, E. 247 
Njorth 132 

Nordwestblock 85, 274 
Norsun Tepe 30 
Norwegian 84 
Nostratic 159 
Novoalekseevka 220 

Novodanilovka culture 197, 202-3, 205, 218, 
234-5; 9J> 99i 'oo, 101, ro2 y m 

oak 13, 112, 115, 187; 2 
October Equus 136 
Odinn 132, 140 
Okunevo culture 223, 263 
Old Church Slavonic 76; 46 
Oscan 90, 94 
Osco-Umbrian 90, 93 
Ossetes 49 

Painted Grey Ware culture 46; 28 

Palaic 25-7, 32, 178 

Panjabi 36 

Parjanyas 129 

Parnassos 179 

Parsons, J. 9-1 1, 143; j, 4 

Parthian 51, 53 

Pashto 48 

Pazhok 70 

Pelasgian 68, 69, 71; 41 
Penka, K. 268 
Penkov culture 78; 49 
Perkunas 129; 55 
Persian 12, 48-9, 54-5; j, 3 i 
Pcrun 1 29; 55 
Petrenko, A. 192 
Petro-Svistunovo 202 

Phrygian 30-35, 58, 61, 63, 108, 153-5, 178, 

182, 222, 241, 264, 273; 79; 3 
Pianello-Timmari horizon 92 
Picene 90-2; 63 
Pictish 106 

pig in, 118-19, 133, 135, 149, 187, 189^91, 
»93, »99> 203, 211, 217, 228, 237, 250, 274 
Piggott, S. 42 
Pindar 132 
Pisani, V. 157 

plough 117, 119, 121, 126, 158, 163, 179, 218, 

Poesche, T. 268 
Poliakov, L. 266 
Polish 18, 77; j, 46 
Pollux 132, 135 
Portuguese 89, 265 
Poseidon 129 
Potekhina, I. 201 
Potemkhina, T. 225-6 

Prague culture 78, 82; 49 
Pre-Caspian culture 206-7 
Prendi, F. 234 

priest deities (first function) 13 1-3, 135, 

138-9, 142 
Prussian 81, 83 
Przewor culture 79-80; 5/ 
Ptolemy 80, 83, 96 
Puhvel, J. 135-0, 140 
Pulgram, E. 257 

quern 119, 199, 212, 218 
Quirinus 132 

Raetic 89 

Rakushchechny Yar 191 
Rask, R. 13-14 
Reichelt, H. 122 
Remedello culture 93 
Remus 140 

Renfrew, C. 7-8, 45, 65, 166-8, 1 77-181, 

242-3, 273-5 
right-left dichotomy 140, 205, 221, 230, 244, 

247, 254 
Rinaldone culture 93-4; 66 
Ripley, J. 66 

rivers names: Baltic 83-4; 54; Finno-Ugric 84; 

Germanic 85; Indo-European 84, 157, 276; 

Iranian 43, 77; 48; Nordwestblock 85; 

Slavic 80-1, 83-4; 52 
Roman, P. 260 

Romance languages 77, 89; /, 2 
Romanian 89 

Roman 10, 72, 74, 88-9, 129, 132, 140, 266, 

Romany (Gypsy) 36, 269 
Romulus 140 

Rossen cutture 25 1 , 253, 257 
Russian 18, 76-7, 265; j, 46 
Rzhevo 234 

Sabine War 139 

Saka 49, 53, 58, 61 

Salcuja culture 234 

salmon 58, 113, 116, 160; 84 

Samara culture 197, 206-7, 210, 226, 262; 110 

Samnites 90 

Sanskrit 8, 12, 16, 36, 44, 56, 82, 157; 2/ 
Sapalli Tepe 231 
Sarasvati 132 

Sarmatian 48, 53, 77, 79, 168, 229; 6 

Sautramani 133 

Sayce, A. H. 143 

Scaliger, J. 9-10 

Scharfe, H. 125 

Schleicher, A. 14-20, in; 6, 7 

Schmid, W. P. 84, 157, 276 

Schmidt, J. 19-20, 182; 8 y 11 

Schrader, O. 269 

Sclavini 76-7 

Scythes 130, 139; 5 

Scythian 10, 48, 50, 53, 79-80, 83, 130, 132, 

139, 228; 5 
Sedov, V. 79 
Selenkahiyeh 41 

Semites 10, 26, 34, 63 4, 69, 118, 149 50, 

163, 259, 272 
Serbian 77; 46 

Seroglazovo culture 192, 206, 210; 8g 
Sezzhee 206, 220, 256; 122; 23 
Shah Tepe 39-40 
Shahr-i Sokhta 44 













sheep 16-17, m, 117, 133, 135, 191, 193, 
199, 203, 2ii 12, 217-18, 220, 223, 227-8, 
240, 242, 250 

Sherratt, A. 126, 179, 242, 260 

Shilov, V. 212 

Shmagliy, N. 239 

Sialk 49 

Siculan 90 

Sihler, A. 125 

silver 47, i2i, 127, 150, 158, 164, 179, 205, 

218, 223, 237, 240 
Sintashta 53, 228 

sky god 128-9, ! 38, 141, 184, 219; 19 
Slavic 8, 10, 18, 58, 73, 76-82, 108, 145, 150, 

i53"5> 157, 182, 223, 243, 255-6; 46, 47, 

S0~3\ 15 
Slovakian 77 
Slovenian 77 
Sogdian 49, 51, 53, 61 
Spanish 89, 265; 3 

Sredny Stog culture 197-203, 205, 207, 
210-n, 213, 215, 218, 220, 234, 248, 
252 3> 255-6, 262; 9/, 93, 94, 95, 96, 97, 
98, in, 26 
Srubnaya culture 50, 52-3, 228-9; 7 26 
stela 204, 219, 22i, 250-1; 119, 120, 121 
Sumerian 26, 41, 44, 64, 121, 130, 147, 150, 

163, 270, 
sun god 129, 184 
suovetaurilia 133 

Sursko-Dnieper culture 190, 205; 86, 87, 88 

suttee 93, 184, 250 

Suvodol 234 

Suvorovo 234 

Sveshnikov, I. 248, 251 

Swedish 84; 3, 57 

Swiderian-Kunda culture 188 

Tacitus 80, 83, 85, 124, 140, 266 
Tal-i IbHs 41 
Taranis 139 

Tartessan 105, 147, 181; 70 
Taylor, I. 268 

Tazabagyab culture 230, 244 

Tclegin, D. 191, 200-1, 205, 215, 253, 256 

Tepecik 30 

Tepe Giyan 39 

Tepe Yahya 44 

Terramare culture 93 

Teutones 85 
Teutates 129 
Thieme, P. 128 
Thor 129, 132; 55 

Thracian 10, 14, 32, 34-5, 72-3, 76, 108, 153, 

182, 241, 263- 4; 12 
Thraetaona 137 
thunder god 122, 141, 219 
Tilak, C. 143, 269, 277 
Tocharian 56-63, 113, 15 1-5, 180, 182, 223, 

226, 256, 261-3, 274; 3, 36, 37, 38 
Todorova, H. 234 
Tomb of the Widow 93; 66 
Toshchev, G. 239 

TRB culture 162-4, 182-3, 186, 217 18, 

248-53, 255; 142 
trees (Indo-European) 113-17, 180, 187, 216 
Trctyakov, V. 248, 256 
Trialcti 232 

Tripolye culture 190, 194, 196-8, 201, 205, 

233-8, 242-3, 248-9, 251, 258, 262; 91, 92 
Trita Aptya 137 
trittua 135 
trout 116^17, 160 

Troy 24, 28^9, 31, 33, 239, 241, 264, 266 
Trubachev, O. 80, 151 
Trzciniec culture 81, 84; 53 
Tuisto 140 

Tulkhar 53, 230, 244; 34, 35 
Tumek-Kichidzhik 226 
Tureng Tepe 40 
Turfan 56, 50-60 

Turkish 10, 61, 63-4, 73, 147, 151-2, 223, 

261, 270; 4, 81 
Tyr 132, 140 

Ubagan I 226 
Uch-Tepe 29, 231, 233 
Ukrainian 18, 77 
Umbrian 90, 94; 6/ 

Uralic 147, 152, 159, 179, 207, 217, 268; 82 
Urartian 34-5, 50, 55, 63, 64, 163, 178 
Urnfield culture 92, 106 
Usatovo culture 237-9; 132, 133, 134 

Vanir 139 

Varuna 37, 42, 13 1-3, 140 

Vasiliev, I. 206, 210, 215, 277 

vedas 37, 47, 52, 54, in, 131, 133, 229; 23 

vehicles 39, 41, 62, 117, 121, 126-7, 158, 
179, 183- 4, 195, 2U-i4» 218, 223, 228 
231-3, 242, 244, 252, 262, 270, 275-6; 
114; 28 

Vencti (Italy) 91 

Veneti (Slavs) 76-77 

Venctic 91, 153; 64 

Verkhnaya Alabuga 226 

Villanovan culture 92-3 

Vinca culture 233 

Vinogradov, A. 227 

Vinogradovka 213 

Virchow, R. 269 

Vodhine 70 

Voroshilovgrad 202 

Vucedol 184 

I2 4"5> 135, 179, 18 

war band 38, 1 10 11 

warrior deities (second function) 131-3, 

135-7, >39, 141-2, 184, 219 
Welsh 10, 18, 96; 3 
Werner, J. 79 
Winn, M. 232 
W r islanski, T. 251 

Wolf 108, IIO, Il6, I57, 187, 200, 212 

von Wolzogen, H. 24 

wool 118, 121, 126, 158, 179, 218, 242 

Wu-Sun 50-60 

Yaghnobis 49 

Yama (god) 129, 140; 99 

Yama (site) 202 

Yamnaya culture 197, 203-5, 210-15, 220 
223, 225-6, 228, 236, 241, 246, 249, 251 
262; H2, 113, 114, 7/5, //6, 7/7, 118, 

Yiddish 269 

Yima 140 

Ymir 140 

yoke 117, 121, 126, 158, 179, 242, 248-9 
Yorgan Tepe 38 
Young, T. 14 
Young, T. C. 50 
Yueh-Chih 60 

Zarathustra 42 3, 52, 271; 33; 4 
Zarubinets culture 79- 80; 31 
Zbenovich, V. 236 
Zeus 128, 130; 79